From ketancmaheshwari at gmail.com Mon Apr 1 15:50:23 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Mon, 1 Apr 2013 15:50:23 -0500 Subject: [Swift-devel] unitialized error when swift app returning a non-file type Message-ID: Hi, Wondering if this is an expected behavior. I intend a Swift app to return a variable which I then receive in a Swift variable. This variable is uninitialized since initializing it would close it out. Here is a minuscule example: app (string b) s_app(string a){ str_app a b; } string[] b; foreach i in [0:10]{ b[i]= s_app("hello"); } The s_app links to a bash script as follows: #!/bin/bash $2=$1 On running this: $ swift str.swift -tc.file tc -sites.file sites.xml Could not start execution Uninitalized variable: b The detailed exception from the log is: 2013-04-01 15:47:33,755-0500 DEBUG Loader Detailed exception: org.griphyn.vdl.karajan.CompilationException: Uninitalized variable: b at org.griphyn.vdl.engine.VariableScope.undeclare(VariableScope.java:1143) at org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1119) at org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) at org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) at org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1085) at org.griphyn.vdl.engine.Karajan.program(Karajan.java:316) at org.griphyn.vdl.engine.Karajan.compile(Karajan.java:137) at org.griphyn.vdl.karajan.Loader.compile(Loader.java:351) at org.griphyn.vdl.karajan.Loader.main(Loader.java:172) Thanks for any help. -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From hategan at mcs.anl.gov Mon Apr 1 16:03:44 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Mon, 01 Apr 2013 14:03:44 -0700 Subject: [Swift-devel] unitialized error when swift app returning a non-file type In-Reply-To: References: Message-ID: <1364850224.23092.0.camel@echo> It's a bug. Working on it. On Mon, 2013-04-01 at 15:50 -0500, Ketan Maheshwari wrote: > Hi, > > Wondering if this is an expected behavior. I intend a Swift app to return a > variable which I then receive in a Swift variable. This variable is > uninitialized since initializing it would close it out. > > Here is a minuscule example: > > app (string b) s_app(string a){ > str_app a b; > } > > string[] b; > > foreach i in [0:10]{ > b[i]= s_app("hello"); > } > > The s_app links to a bash script as follows: > > #!/bin/bash > > $2=$1 > > On running this: > > $ swift str.swift -tc.file tc -sites.file sites.xml > Could not start execution > Uninitalized variable: b > > > The detailed exception from the log is: > 2013-04-01 15:47:33,755-0500 DEBUG Loader Detailed exception: > org.griphyn.vdl.karajan.CompilationException: Uninitalized variable: b > at > org.griphyn.vdl.engine.VariableScope.undeclare(VariableScope.java:1143) > at > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1119) > at > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > at > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > at > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1085) > at org.griphyn.vdl.engine.Karajan.program(Karajan.java:316) > at org.griphyn.vdl.engine.Karajan.compile(Karajan.java:137) > at org.griphyn.vdl.karajan.Loader.compile(Loader.java:351) > at org.griphyn.vdl.karajan.Loader.main(Loader.java:172) > > Thanks for any help. > > Hi, > > > Wondering if this is an expected behavior. I intend a Swift app to > return a variable which I then receive in a Swift variable. This > variable is uninitialized since initializing it would close it out. > > > Here is a minuscule example: > > > app (string b) s_app(string a){ > str_app a b; > } > > > string[] b; > > > foreach i in [0:10]{ > b[i]= s_app("hello"); > } > > > The s_app links to a bash script as follows: > > > #!/bin/bash > > > $2=$1 > > > On running this: > > > $ swift str.swift -tc.file tc -sites.file sites.xml > Could not start execution > Uninitalized variable: b > > > > > The detailed exception from the log is: > 2013-04-01 15:47:33,755-0500 DEBUG Loader Detailed exception: > org.griphyn.vdl.karajan.CompilationException: Uninitalized variable: b > at > org.griphyn.vdl.engine.VariableScope.undeclare(VariableScope.java:1143) > at > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1119) > at > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > at > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > at > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1085) > at org.griphyn.vdl.engine.Karajan.program(Karajan.java:316) > at org.griphyn.vdl.engine.Karajan.compile(Karajan.java:137) > at org.griphyn.vdl.karajan.Loader.compile(Loader.java:351) > at org.griphyn.vdl.karajan.Loader.main(Loader.java:172) > > > Thanks for any help. > > > -- > Ketan > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel From hategan at mcs.anl.gov Mon Apr 1 16:10:39 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Mon, 01 Apr 2013 14:10:39 -0700 Subject: [Swift-devel] unitialized error when swift app returning a non-file type In-Reply-To: <1364850224.23092.0.camel@echo> References: <1364850224.23092.0.camel@echo> Message-ID: <1364850639.23092.2.camel@echo> Wait, it's no bug. It's a poor error message. This refers to the 'b' that is a return parameter in s_app. You are using it as an input parameter on the command line. Mihael On Mon, 2013-04-01 at 14:03 -0700, Mihael Hategan wrote: > It's a bug. Working on it. > > On Mon, 2013-04-01 at 15:50 -0500, Ketan Maheshwari wrote: > > Hi, > > > > Wondering if this is an expected behavior. I intend a Swift app to return a > > variable which I then receive in a Swift variable. This variable is > > uninitialized since initializing it would close it out. > > > > Here is a minuscule example: > > > > app (string b) s_app(string a){ > > str_app a b; > > } > > > > string[] b; > > > > foreach i in [0:10]{ > > b[i]= s_app("hello"); > > } > > > > The s_app links to a bash script as follows: > > > > #!/bin/bash > > > > $2=$1 > > > > On running this: > > > > $ swift str.swift -tc.file tc -sites.file sites.xml > > Could not start execution > > Uninitalized variable: b > > > > > > The detailed exception from the log is: > > 2013-04-01 15:47:33,755-0500 DEBUG Loader Detailed exception: > > org.griphyn.vdl.karajan.CompilationException: Uninitalized variable: b > > at > > org.griphyn.vdl.engine.VariableScope.undeclare(VariableScope.java:1143) > > at > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1119) > > at > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > > at > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > > at > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1085) > > at org.griphyn.vdl.engine.Karajan.program(Karajan.java:316) > > at org.griphyn.vdl.engine.Karajan.compile(Karajan.java:137) > > at org.griphyn.vdl.karajan.Loader.compile(Loader.java:351) > > at org.griphyn.vdl.karajan.Loader.main(Loader.java:172) > > > > Thanks for any help. > > > > Hi, > > > > > > Wondering if this is an expected behavior. I intend a Swift app to > > return a variable which I then receive in a Swift variable. This > > variable is uninitialized since initializing it would close it out. > > > > > > Here is a minuscule example: > > > > > > app (string b) s_app(string a){ > > str_app a b; > > } > > > > > > string[] b; > > > > > > foreach i in [0:10]{ > > b[i]= s_app("hello"); > > } > > > > > > The s_app links to a bash script as follows: > > > > > > #!/bin/bash > > > > > > $2=$1 > > > > > > On running this: > > > > > > $ swift str.swift -tc.file tc -sites.file sites.xml > > Could not start execution > > Uninitalized variable: b > > > > > > > > > > The detailed exception from the log is: > > 2013-04-01 15:47:33,755-0500 DEBUG Loader Detailed exception: > > org.griphyn.vdl.karajan.CompilationException: Uninitalized variable: b > > at > > org.griphyn.vdl.engine.VariableScope.undeclare(VariableScope.java:1143) > > at > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1119) > > at > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > > at > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > > at > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1085) > > at org.griphyn.vdl.engine.Karajan.program(Karajan.java:316) > > at org.griphyn.vdl.engine.Karajan.compile(Karajan.java:137) > > at org.griphyn.vdl.karajan.Loader.compile(Loader.java:351) > > at org.griphyn.vdl.karajan.Loader.main(Loader.java:172) > > > > > > Thanks for any help. > > > > > > -- > > Ketan > > > > > > _______________________________________________ > > Swift-devel mailing list > > Swift-devel at ci.uchicago.edu > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel From ketancmaheshwari at gmail.com Mon Apr 1 16:18:52 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Mon, 1 Apr 2013 16:18:52 -0500 Subject: [Swift-devel] unitialized error when swift app returning a non-file type In-Reply-To: <1364850639.23092.2.camel@echo> References: <1364850224.23092.0.camel@echo> <1364850639.23092.2.camel@echo> Message-ID: To my understanding, it has to be used in the commandline so that Swift knows where to put its value. Here is a variant which does not use it in command line but nothing is returned too: app (string _b) s_app(string a){ str_app a; } string[] b; foreach i in [0:10]{ b[i]= s_app("hello"); tracef("%s\n", b[i]); } $ swift str.swift -tc.file tc -sites.file sites.xml Swift trunk swift-r6410 cog-r3648 RunID: 20130401-1617-4k42kvp5 Progress: time: Mon, 01 Apr 2013 16:17:58 -0500 null null Progress: time: Mon, 01 Apr 2013 16:17:59 -0500 Selecting site:8 Stage in:1 Finished successfully:2 null null Progress: time: Mon, 01 Apr 2013 16:18:01 -0500 Selecting site:6 Stage in:1 Finished successfully:4 null null Progress: time: Mon, 01 Apr 2013 16:18:02 -0500 Selecting site:4 Stage in:1 Finished successfully:6 null null Progress: time: Mon, 01 Apr 2013 16:18:03 -0500 Selecting site:2 Stage in:1 Finished successfully:8 null null Progress: time: Mon, 01 Apr 2013 16:18:05 -0500 Stage in:1 Finished successfully:10 null Final status: Mon, 01 Apr 2013 16:18:05 -0500 Finished successfully:11 On Mon, Apr 1, 2013 at 4:10 PM, Mihael Hategan wrote: > Wait, it's no bug. > > It's a poor error message. This refers to the 'b' that is a return > parameter in s_app. You are using it as an input parameter on the > command line. > > Mihael > > On Mon, 2013-04-01 at 14:03 -0700, Mihael Hategan wrote: > > It's a bug. Working on it. > > > > On Mon, 2013-04-01 at 15:50 -0500, Ketan Maheshwari wrote: > > > Hi, > > > > > > Wondering if this is an expected behavior. I intend a Swift app to > return a > > > variable which I then receive in a Swift variable. This variable is > > > uninitialized since initializing it would close it out. > > > > > > Here is a minuscule example: > > > > > > app (string b) s_app(string a){ > > > str_app a b; > > > } > > > > > > string[] b; > > > > > > foreach i in [0:10]{ > > > b[i]= s_app("hello"); > > > } > > > > > > The s_app links to a bash script as follows: > > > > > > #!/bin/bash > > > > > > $2=$1 > > > > > > On running this: > > > > > > $ swift str.swift -tc.file tc -sites.file sites.xml > > > Could not start execution > > > Uninitalized variable: b > > > > > > > > > The detailed exception from the log is: > > > 2013-04-01 15:47:33,755-0500 DEBUG Loader Detailed exception: > > > org.griphyn.vdl.karajan.CompilationException: Uninitalized variable: b > > > at > > > org.griphyn.vdl.engine.VariableScope.undeclare(VariableScope.java:1143) > > > at > > > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1119) > > > at > > > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > > > at > > > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > > > at > > > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1085) > > > at org.griphyn.vdl.engine.Karajan.program(Karajan.java:316) > > > at org.griphyn.vdl.engine.Karajan.compile(Karajan.java:137) > > > at org.griphyn.vdl.karajan.Loader.compile(Loader.java:351) > > > at org.griphyn.vdl.karajan.Loader.main(Loader.java:172) > > > > > > Thanks for any help. > > > > > > Hi, > > > > > > > > > Wondering if this is an expected behavior. I intend a Swift app to > > > return a variable which I then receive in a Swift variable. This > > > variable is uninitialized since initializing it would close it out. > > > > > > > > > Here is a minuscule example: > > > > > > > > > app (string b) s_app(string a){ > > > str_app a b; > > > } > > > > > > > > > string[] b; > > > > > > > > > foreach i in [0:10]{ > > > b[i]= s_app("hello"); > > > } > > > > > > > > > The s_app links to a bash script as follows: > > > > > > > > > #!/bin/bash > > > > > > > > > $2=$1 > > > > > > > > > On running this: > > > > > > > > > $ swift str.swift -tc.file tc -sites.file sites.xml > > > Could not start execution > > > Uninitalized variable: b > > > > > > > > > > > > > > > The detailed exception from the log is: > > > 2013-04-01 15:47:33,755-0500 DEBUG Loader Detailed exception: > > > org.griphyn.vdl.karajan.CompilationException: Uninitalized variable: b > > > at > > > org.griphyn.vdl.engine.VariableScope.undeclare(VariableScope.java:1143) > > > at > > > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1119) > > > at > > > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > > > at > > > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1115) > > > at > > > > org.griphyn.vdl.engine.VariableScope.analyzeWriters(VariableScope.java:1085) > > > at org.griphyn.vdl.engine.Karajan.program(Karajan.java:316) > > > at org.griphyn.vdl.engine.Karajan.compile(Karajan.java:137) > > > at org.griphyn.vdl.karajan.Loader.compile(Loader.java:351) > > > at org.griphyn.vdl.karajan.Loader.main(Loader.java:172) > > > > > > > > > Thanks for any help. > > > > > > > > > -- > > > Ketan > > > > > > > > > _______________________________________________ > > > Swift-devel mailing list > > > Swift-devel at ci.uchicago.edu > > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > > > > _______________________________________________ > > Swift-devel mailing list > > Swift-devel at ci.uchicago.edu > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From hategan at mcs.anl.gov Mon Apr 1 16:23:33 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Mon, 01 Apr 2013 14:23:33 -0700 Subject: [Swift-devel] unitialized error when swift app returning a non-file type In-Reply-To: References: <1364850224.23092.0.camel@echo> <1364850639.23092.2.camel@echo> Message-ID: <1364851413.23590.2.camel@echo> On Mon, 2013-04-01 at 16:18 -0500, Ketan Maheshwari wrote: > To my understanding, it has to be used in the commandline so that Swift > knows where to put its value. I'm not sure what you mean there, or where you got that information, but when you say app ... { appname x; } that means invoke 'appname' with the value of x as a parameter on the command line. There is no such value if x is a return parameter. Mihael From ketancmaheshwari at gmail.com Mon Apr 1 16:33:26 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Mon, 1 Apr 2013 16:33:26 -0500 Subject: [Swift-devel] unitialized error when swift app returning a non-file type In-Reply-To: <1364851413.23590.2.camel@echo> References: <1364850224.23092.0.camel@echo> <1364850639.23092.2.camel@echo> <1364851413.23590.2.camel@echo> Message-ID: Consider the following variant of above example: type file; app (file _b) s_app(string a){ str_app a @_b; } file b[]; foreach i in [0:10]{ b[i]= s_app("hello"); } And the corresponding bash script that app links to: #!/bin/bash echo "$1">$2 This variant exhibits the same pattern as the previous one except that b now is a file type variable. The original example which did not work involved b as a non-file type variable (string). Since this one works, I understood the one with other types should also work. On Mon, Apr 1, 2013 at 4:23 PM, Mihael Hategan wrote: > On Mon, 2013-04-01 at 16:18 -0500, Ketan Maheshwari wrote: > > To my understanding, it has to be used in the commandline so that Swift > > knows where to put its value. > > I'm not sure what you mean there, or where you got that information, but > when you say > > app ... { > appname x; > } > > that means invoke 'appname' with the value of x as a parameter on the > command line. There is no such value if x is a return parameter. > > Mihael > > > -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From hategan at mcs.anl.gov Mon Apr 1 16:49:58 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Mon, 01 Apr 2013 14:49:58 -0700 Subject: [Swift-devel] unitialized error when swift app returning a non-file type In-Reply-To: References: <1364850224.23092.0.camel@echo> <1364850639.23092.2.camel@echo> <1364851413.23590.2.camel@echo> Message-ID: <1364852998.23816.7.camel@echo> What do you think should appear on the command line for echo in the following case: app (string r) f() { echo r; } string s; s = f(); ? Mihael On Mon, 2013-04-01 at 16:33 -0500, Ketan Maheshwari wrote: > Consider the following variant of above example: > > type file; > > app (file _b) s_app(string a){ > str_app a @_b; > } > > file b[]; > > foreach i in [0:10]{ > b[i]= s_app("hello"); > } > > And the corresponding bash script that app links to: > > #!/bin/bash > > echo "$1">$2 > > This variant exhibits the same pattern as the previous one except that b > now is a file type variable. The original example which did not work > involved b as a non-file type variable (string). > > Since this one works, I understood the one with other types should also > work. > > > > On Mon, Apr 1, 2013 at 4:23 PM, Mihael Hategan wrote: > > > On Mon, 2013-04-01 at 16:18 -0500, Ketan Maheshwari wrote: > > > To my understanding, it has to be used in the commandline so that Swift > > > knows where to put its value. > > > > I'm not sure what you mean there, or where you got that information, but > > when you say > > > > app ... { > > appname x; > > } > > > > that means invoke 'appname' with the value of x as a parameter on the > > command line. There is no such value if x is a return parameter. > > > > Mihael > > > > > > > > > Consider the following variant of above example: > > > type file; > > > app (file _b) s_app(string a){ > str_app a @_b; > } > > > file b[]; > > > foreach i in [0:10]{ > b[i]= s_app("hello"); > } > > > And the corresponding bash script that app links to: > > > #!/bin/bash > > > echo "$1">$2 > > > This variant exhibits the same pattern as the previous one except that > b now is a file type variable. The original example which did not work > involved b as a non-file type variable (string). > > > Since this one works, I understood the one with other types should > also work. > > > > > On Mon, Apr 1, 2013 at 4:23 PM, Mihael Hategan > wrote: > On Mon, 2013-04-01 at 16:18 -0500, Ketan Maheshwari wrote: > > To my understanding, it has to be used in the commandline so > that Swift > > knows where to put its value. > > > I'm not sure what you mean there, or where you got that > information, but > when you say > > app ... { > appname x; > } > > that means invoke 'appname' with the value of x as a parameter > on the > command line. There is no such value if x is a return > parameter. > > Mihael > > > > > > > -- > Ketan > > From ketancmaheshwari at gmail.com Mon Apr 1 17:49:36 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Mon, 1 Apr 2013 17:49:36 -0500 Subject: [Swift-devel] cache already contains error Message-ID: Hi, I am running into the "cache already contains" error when using a nested loop with file mappers. Here is a simple reproduction of the issue with a nested loop variant of catsn.swift: type file; app (file o) cat (file i){ cat @i stdout=@o; } #file out[]; #file out[]; file out[]; foreach i in [0:1] { foreach j in [0:1]{ file data<"data.txt">; out[j] = cat(data); } } It runs into the cache error after completing few tasks successfully: $ swift catsn.swift Swift trunk swift-r6410 cog-r3648 RunID: 20130401-1745-7khkyrqc Progress: time: Mon, 01 Apr 2013 17:45:59 -0500 Progress: time: Mon, 01 Apr 2013 17:46:00 -0500 Selecting site:1 Active:1 Finished successfully:2 Execution failed: Exception in cat: Arguments: [data.txt] Host: localhost Directory: catsn-20130401-1745-7khkyrqc/jobs/y/cat-yzf9fg7l Caused by: The cache already contains localhost:catsn-20130401-1745-7khkyrqc/shared/outdir/f.0000.out. cat, catsn.swift, line 14 The cause, I think is that the nested loop triggers the same series of random sequences in mappers code which collides. Both, the simple and the concurrent mappers fail with same message. Does anyone know of a workaround? Thanks, -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From wilde at mcs.anl.gov Mon Apr 1 17:59:15 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Mon, 1 Apr 2013 17:59:15 -0500 (CDT) Subject: [Swift-devel] cache already contains error In-Reply-To: Message-ID: <932458407.3332912.1364857155365.JavaMail.root@mcs.anl.gov> I think you need to make out 2-dimensional. Your script is going to evaluate "out[j] = cat(data)" for both i=0 and i=1. The second of those evaluations is probably encountering the "cache already contains" for j=0. If it didnt hit that (ie if you used the concurrent mapper) you'd likely then get an error that out[0] is already set. - Mike ----- Original Message ----- > From: "Ketan Maheshwari" > To: "Swift Devel" > Sent: Monday, April 1, 2013 5:49:36 PM > Subject: [Swift-devel] cache already contains error > > > > > Hi, > > I am running into the "cache already contains" error when using a > nested loop with file mappers. Here is a simple reproduction of the > issue with a nested loop variant of catsn.swift: > > > > type file; > app (file o) cat (file i){ > cat @i stdout=@o; > } > > > #file out[]; > #file out[] prefix="f.",suffix=".out">; > file out[] prefix="f.",suffix=".out">; > > > foreach i in [0:1] { > foreach j in [0:1]{ > file data<"data.txt">; > out[j] = cat(data); > } > } > > > It runs into the cache error after completing few tasks successfully: > > $ swift catsn.swift > Swift trunk swift-r6410 cog-r3648 > > > RunID: 20130401-1745-7khkyrqc > Progress: time: Mon, 01 Apr 2013 17:45:59 -0500 > Progress: time: Mon, 01 Apr 2013 17:46:00 -0500 Selecting site:1 > Active:1 Finished successfully:2 > Execution failed: > Exception in cat: > Arguments: [data.txt] > Host: localhost > Directory: catsn-20130401-1745-7khkyrqc/jobs/y/cat-yzf9fg7l > Caused by: > The cache already contains > localhost:catsn-20130401-1745-7khkyrqc/shared/outdir/f.0000.out. > cat, catsn.swift, line 14 > > > The cause, I think is that the nested loop triggers the same series > of random sequences in mappers code which collides. Both, the simple > and the concurrent mappers fail with same message. > > > Does anyone know of a workaround? > > > Thanks, > -- > Ketan > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > From ketancmaheshwari at gmail.com Mon Apr 1 21:02:24 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Mon, 1 Apr 2013 21:02:24 -0500 Subject: [Swift-devel] cache already contains error In-Reply-To: <932458407.3332912.1364857155365.JavaMail.root@mcs.anl.gov> References: <932458407.3332912.1364857155365.JavaMail.root@mcs.anl.gov> Message-ID: Thanks Mike, that fixed the cache issue. However, now I am seeing an unusual behavior from my Swift run: The ampl run crashes after completing a fixed number of jobs (35 to be precise). Some diagnostics: -- It runs to completion when I do a Swift resume. Once again only the next 35 jobs complete successfully. On a next resume the rest of them complete. -- Runs outside of Swift with a bash for-loop using the same parameters as in Swift script. -- A catsn script of similar parameters runs to completion without any failures. So, nothing seem to be wrong with the OS parameters. I am using a single MCS workstation, no provider staging, no coasters. The error message is: Caused by: File not found: /nfs2/ketan/powergridapps/swiftscripts/swift.work/inference-20130401-2043-78i7o5m3/shared/outdir/out_l0000_0000.0010.out Which is reflected in the logs as well as in the workdir's info files. Has anyone seen this kind of behavior? Any remedial suggestions? Thanks, Ketan On Mon, Apr 1, 2013 at 5:59 PM, Michael Wilde wrote: > I think you need to make out 2-dimensional. > > Your script is going to evaluate "out[j] = cat(data)" for both i=0 and i=1. > > The second of those evaluations is probably encountering the "cache > already contains" for j=0. > > If it didnt hit that (ie if you used the concurrent mapper) you'd likely > then get an error that out[0] is already set. > > - Mike > > > ----- Original Message ----- > > From: "Ketan Maheshwari" > > To: "Swift Devel" > > Sent: Monday, April 1, 2013 5:49:36 PM > > Subject: [Swift-devel] cache already contains error > > > > > > > > > > Hi, > > > > I am running into the "cache already contains" error when using a > > nested loop with file mappers. Here is a simple reproduction of the > > issue with a nested loop variant of catsn.swift: > > > > > > > > type file; > > app (file o) cat (file i){ > > cat @i stdout=@o; > > } > > > > > > #file out[]; > > #file out[] > prefix="f.",suffix=".out">; > > file out[] > prefix="f.",suffix=".out">; > > > > > > foreach i in [0:1] { > > foreach j in [0:1]{ > > file data<"data.txt">; > > out[j] = cat(data); > > } > > } > > > > > > It runs into the cache error after completing few tasks successfully: > > > > $ swift catsn.swift > > Swift trunk swift-r6410 cog-r3648 > > > > > > RunID: 20130401-1745-7khkyrqc > > Progress: time: Mon, 01 Apr 2013 17:45:59 -0500 > > Progress: time: Mon, 01 Apr 2013 17:46:00 -0500 Selecting site:1 > > Active:1 Finished successfully:2 > > Execution failed: > > Exception in cat: > > Arguments: [data.txt] > > Host: localhost > > Directory: catsn-20130401-1745-7khkyrqc/jobs/y/cat-yzf9fg7l > > Caused by: > > The cache already contains > > localhost:catsn-20130401-1745-7khkyrqc/shared/outdir/f.0000.out. > > cat, catsn.swift, line 14 > > > > > > The cause, I think is that the nested loop triggers the same series > > of random sequences in mappers code which collides. Both, the simple > > and the concurrent mappers fail with same message. > > > > > > Does anyone know of a workaround? > > > > > > Thanks, > > -- > > Ketan > > > > > > _______________________________________________ > > Swift-devel mailing list > > Swift-devel at ci.uchicago.edu > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From ketancmaheshwari at gmail.com Mon Apr 1 21:26:17 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Mon, 1 Apr 2013 21:26:17 -0500 Subject: [Swift-devel] unitialized error when swift app returning a non-file type In-Reply-To: <1364852998.23816.7.camel@echo> References: <1364850224.23092.0.camel@echo> <1364850639.23092.2.camel@echo> <1364851413.23590.2.camel@echo> <1364852998.23816.7.camel@echo> Message-ID: I was thinking Swift would take a placeholder for 'r' which would be a command line arg and would expect the placeholder be filled with value in the app call. However, I agree that in this example it would be counter intuitive and perhaps unsuitable. Does this mean that a Swift app can only return a file type? On Mon, Apr 1, 2013 at 4:49 PM, Mihael Hategan wrote: > What do you think should appear on the command line for echo in the > following case: > > app (string r) f() { > echo r; > } > > string s; > s = f(); > ? > > Mihael > > On Mon, 2013-04-01 at 16:33 -0500, Ketan Maheshwari wrote: > > Consider the following variant of above example: > > > > type file; > > > > app (file _b) s_app(string a){ > > str_app a @_b; > > } > > > > file b[]; > > > > foreach i in [0:10]{ > > b[i]= s_app("hello"); > > } > > > > And the corresponding bash script that app links to: > > > > #!/bin/bash > > > > echo "$1">$2 > > > > This variant exhibits the same pattern as the previous one except that b > > now is a file type variable. The original example which did not work > > involved b as a non-file type variable (string). > > > > Since this one works, I understood the one with other types should also > > work. > > > > > > > > On Mon, Apr 1, 2013 at 4:23 PM, Mihael Hategan > wrote: > > > > > On Mon, 2013-04-01 at 16:18 -0500, Ketan Maheshwari wrote: > > > > To my understanding, it has to be used in the commandline so that > Swift > > > > knows where to put its value. > > > > > > I'm not sure what you mean there, or where you got that information, > but > > > when you say > > > > > > app ... { > > > appname x; > > > } > > > > > > that means invoke 'appname' with the value of x as a parameter on the > > > command line. There is no such value if x is a return parameter. > > > > > > Mihael > > > > > > > > > > > > > > > Consider the following variant of above example: > > > > > > type file; > > > > > > app (file _b) s_app(string a){ > > str_app a @_b; > > } > > > > > > file b[]; > > > > > > foreach i in [0:10]{ > > b[i]= s_app("hello"); > > } > > > > > > And the corresponding bash script that app links to: > > > > > > #!/bin/bash > > > > > > echo "$1">$2 > > > > > > This variant exhibits the same pattern as the previous one except that > > b now is a file type variable. The original example which did not work > > involved b as a non-file type variable (string). > > > > > > Since this one works, I understood the one with other types should > > also work. > > > > > > > > > > On Mon, Apr 1, 2013 at 4:23 PM, Mihael Hategan > > wrote: > > On Mon, 2013-04-01 at 16:18 -0500, Ketan Maheshwari wrote: > > > To my understanding, it has to be used in the commandline so > > that Swift > > > knows where to put its value. > > > > > > I'm not sure what you mean there, or where you got that > > information, but > > when you say > > > > app ... { > > appname x; > > } > > > > that means invoke 'appname' with the value of x as a parameter > > on the > > command line. There is no such value if x is a return > > parameter. > > > > Mihael > > > > > > > > > > > > > > -- > > Ketan > > > > > > > -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From wilde at mcs.anl.gov Mon Apr 1 22:10:38 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Mon, 1 Apr 2013 22:10:38 -0500 (CDT) Subject: [Swift-devel] unitialized error when swift app returning a non-file type In-Reply-To: Message-ID: <1492114919.3358289.1364872238336.JavaMail.root@mcs.anl.gov> > Does this mean that a Swift app can only return a file type? Yes. If you want to return one or more scalar values from an app, have the app put them in a file in one of the readData() formats, and wrap a Swift function around the app function that does the readData() of the returned file and returns the extracted scalars. - Mike From hategan at mcs.anl.gov Mon Apr 1 22:13:06 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Mon, 01 Apr 2013 20:13:06 -0700 Subject: [Swift-devel] unitialized error when swift app returning a non-file type In-Reply-To: References: <1364850224.23092.0.camel@echo> <1364850639.23092.2.camel@echo> <1364851413.23590.2.camel@echo> <1364852998.23816.7.camel@echo> Message-ID: <1364872386.27726.4.camel@echo> On Mon, 2013-04-01 at 21:26 -0500, Ketan Maheshwari wrote: > I was thinking Swift would take a placeholder for 'r' which would be a > command line arg and would expect the placeholder be filled with value in > the app call. However, I agree that in this example it would be counter > intuitive and perhaps unsuitable. > > Does this mean that a Swift app can only return a file type? Yes. Though I think there should be a way to re-direct stdout and stderr to primitives. Perhaps that could work in a similar way we do it with @filename, so probably close to what you are thinking. From wilde at mcs.anl.gov Mon Apr 1 22:20:27 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Mon, 1 Apr 2013 22:20:27 -0500 (CDT) Subject: [Swift-devel] cache already contains error In-Reply-To: Message-ID: <118967577.3358728.1364872827527.JavaMail.root@mcs.anl.gov> Ketan, Can you post a pointer to your code, the Swift log, and the Swift stdout/err? How are you mapping the file "...outdir/out_l0000_0000.0010.out"? Does 35 correspond to any of the array bounds? Does it fail if you use only one host? (I.e. my first thought was some kind of NSF sync error). Could you try it using local disk and provider staging with the N hosts? - Mike ----- Original Message ----- > From: "Ketan Maheshwari" > To: "Michael Wilde" > Cc: "Swift Devel" > Sent: Monday, April 1, 2013 9:02:24 PM > Subject: Re: [Swift-devel] cache already contains error > > > Thanks Mike, that fixed the cache issue. However, now I am seeing an > unusual behavior from my Swift run: > > > The ampl run crashes after completing a fixed number of jobs (35 to > be precise). > > > Some diagnostics: > > > -- It runs to completion when I do a Swift resume. Once again only > the next 35 jobs complete successfully. On a next resume the rest of > them complete. > > > -- Runs outside of Swift with a bash for-loop using the same > parameters as in Swift script. > > > -- A catsn script of similar parameters runs to completion without > any failures. So, nothing seem to be wrong with the OS parameters. > > > I am using a single MCS workstation, no provider staging, no > coasters. > > > The error message is: > > > Caused by: File not found: > /nfs2/ketan/powergridapps/swiftscripts/swift.work/inference-20130401-2043-78i7o5m3/shared/outdir/out_l0000_0000.0010.out > > > Which is reflected in the logs as well as in the workdir's info > files. > > > Has anyone seen this kind of behavior? Any remedial suggestions? > > > Thanks, > Ketan > > > > On Mon, Apr 1, 2013 at 5:59 PM, Michael Wilde < wilde at mcs.anl.gov > > wrote: > > > I think you need to make out 2-dimensional. > > Your script is going to evaluate "out[j] = cat(data)" for both i=0 > and i=1. > > The second of those evaluations is probably encountering the "cache > already contains" for j=0. > > If it didnt hit that (ie if you used the concurrent mapper) you'd > likely then get an error that out[0] is already set. > > - Mike > > > > > ----- Original Message ----- > > From: "Ketan Maheshwari" < ketancmaheshwari at gmail.com > > > To: "Swift Devel" < swift-devel at ci.uchicago.edu > > > Sent: Monday, April 1, 2013 5:49:36 PM > > Subject: [Swift-devel] cache already contains error > > > > > > > > > > Hi, > > > > I am running into the "cache already contains" error when using a > > nested loop with file mappers. Here is a simple reproduction of the > > issue with a nested loop variant of catsn.swift: > > > > > > > > type file; > > app (file o) cat (file i){ > > cat @i stdout=@o; > > } > > > > > > #file out[]; > > #file out[] > prefix="f.",suffix=".out">; > > file out[] > prefix="f.",suffix=".out">; > > > > > > foreach i in [0:1] { > > foreach j in [0:1]{ > > file data<"data.txt">; > > out[j] = cat(data); > > } > > } > > > > > > It runs into the cache error after completing few tasks > > successfully: > > > > $ swift catsn.swift > > Swift trunk swift-r6410 cog-r3648 > > > > > > RunID: 20130401-1745-7khkyrqc > > Progress: time: Mon, 01 Apr 2013 17:45:59 -0500 > > Progress: time: Mon, 01 Apr 2013 17:46:00 -0500 Selecting site:1 > > Active:1 Finished successfully:2 > > Execution failed: > > Exception in cat: > > Arguments: [data.txt] > > Host: localhost > > Directory: catsn-20130401-1745-7khkyrqc/jobs/y/cat-yzf9fg7l > > Caused by: > > The cache already contains > > localhost:catsn-20130401-1745-7khkyrqc/shared/outdir/f.0000.out. > > cat, catsn.swift, line 14 > > > > > > The cause, I think is that the nested loop triggers the same series > > of random sequences in mappers code which collides. Both, the > > simple > > and the concurrent mappers fail with same message. > > > > > > Does anyone know of a workaround? > > > > > > Thanks, > > -- > > Ketan > > > > > > _______________________________________________ > > Swift-devel mailing list > > Swift-devel at ci.uchicago.edu > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > > > > > -- > Ketan > > From ketancmaheshwari at gmail.com Tue Apr 2 09:43:19 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Tue, 2 Apr 2013 09:43:19 -0500 Subject: [Swift-devel] cache already contains error In-Reply-To: <118967577.3358728.1364872827527.JavaMail.root@mcs.anl.gov> References: <118967577.3358728.1364872827527.JavaMail.root@mcs.anl.gov> Message-ID: Hi Mike, It works with provider staging. So, it does look like an NFS sync issue. Thanks, Ketan On Mon, Apr 1, 2013 at 10:20 PM, Michael Wilde wrote: > Ketan, > > Can you post a pointer to your code, the Swift log, and the Swift > stdout/err? > > How are you mapping the file "...outdir/out_l0000_0000.0010.out"? > > Does 35 correspond to any of the array bounds? > > Does it fail if you use only one host? (I.e. my first thought was some > kind of NSF sync error). > > Could you try it using local disk and provider staging with the N hosts? > > - Mike > > > ----- Original Message ----- > > From: "Ketan Maheshwari" > > To: "Michael Wilde" > > Cc: "Swift Devel" > > Sent: Monday, April 1, 2013 9:02:24 PM > > Subject: Re: [Swift-devel] cache already contains error > > > > > > Thanks Mike, that fixed the cache issue. However, now I am seeing an > > unusual behavior from my Swift run: > > > > > > The ampl run crashes after completing a fixed number of jobs (35 to > > be precise). > > > > > > Some diagnostics: > > > > > > -- It runs to completion when I do a Swift resume. Once again only > > the next 35 jobs complete successfully. On a next resume the rest of > > them complete. > > > > > > -- Runs outside of Swift with a bash for-loop using the same > > parameters as in Swift script. > > > > > > -- A catsn script of similar parameters runs to completion without > > any failures. So, nothing seem to be wrong with the OS parameters. > > > > > > I am using a single MCS workstation, no provider staging, no > > coasters. > > > > > > The error message is: > > > > > > Caused by: File not found: > > > /nfs2/ketan/powergridapps/swiftscripts/swift.work/inference-20130401-2043-78i7o5m3/shared/outdir/out_l0000_0000.0010.out > > > > > > Which is reflected in the logs as well as in the workdir's info > > files. > > > > > > Has anyone seen this kind of behavior? Any remedial suggestions? > > > > > > Thanks, > > Ketan > > > > > > > > On Mon, Apr 1, 2013 at 5:59 PM, Michael Wilde < wilde at mcs.anl.gov > > > wrote: > > > > > > I think you need to make out 2-dimensional. > > > > Your script is going to evaluate "out[j] = cat(data)" for both i=0 > > and i=1. > > > > The second of those evaluations is probably encountering the "cache > > already contains" for j=0. > > > > If it didnt hit that (ie if you used the concurrent mapper) you'd > > likely then get an error that out[0] is already set. > > > > - Mike > > > > > > > > > > ----- Original Message ----- > > > From: "Ketan Maheshwari" < ketancmaheshwari at gmail.com > > > > To: "Swift Devel" < swift-devel at ci.uchicago.edu > > > > Sent: Monday, April 1, 2013 5:49:36 PM > > > Subject: [Swift-devel] cache already contains error > > > > > > > > > > > > > > > Hi, > > > > > > I am running into the "cache already contains" error when using a > > > nested loop with file mappers. Here is a simple reproduction of the > > > issue with a nested loop variant of catsn.swift: > > > > > > > > > > > > type file; > > > app (file o) cat (file i){ > > > cat @i stdout=@o; > > > } > > > > > > > > > #file out[]; > > > #file out[] > > prefix="f.",suffix=".out">; > > > file out[] > > prefix="f.",suffix=".out">; > > > > > > > > > foreach i in [0:1] { > > > foreach j in [0:1]{ > > > file data<"data.txt">; > > > out[j] = cat(data); > > > } > > > } > > > > > > > > > It runs into the cache error after completing few tasks > > > successfully: > > > > > > $ swift catsn.swift > > > Swift trunk swift-r6410 cog-r3648 > > > > > > > > > RunID: 20130401-1745-7khkyrqc > > > Progress: time: Mon, 01 Apr 2013 17:45:59 -0500 > > > Progress: time: Mon, 01 Apr 2013 17:46:00 -0500 Selecting site:1 > > > Active:1 Finished successfully:2 > > > Execution failed: > > > Exception in cat: > > > Arguments: [data.txt] > > > Host: localhost > > > Directory: catsn-20130401-1745-7khkyrqc/jobs/y/cat-yzf9fg7l > > > Caused by: > > > The cache already contains > > > localhost:catsn-20130401-1745-7khkyrqc/shared/outdir/f.0000.out. > > > cat, catsn.swift, line 14 > > > > > > > > > The cause, I think is that the nested loop triggers the same series > > > of random sequences in mappers code which collides. Both, the > > > simple > > > and the concurrent mappers fail with same message. > > > > > > > > > Does anyone know of a workaround? > > > > > > > > > Thanks, > > > -- > > > Ketan > > > > > > > > > _______________________________________________ > > > Swift-devel mailing list > > > Swift-devel at ci.uchicago.edu > > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > > > > > > > > > > > -- > > Ketan > > > > > -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From wilde at mcs.anl.gov Tue Apr 2 10:04:06 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Tue, 2 Apr 2013 10:04:06 -0500 (CDT) Subject: [Swift-devel] cache already contains error In-Reply-To: Message-ID: <33477128.3459312.1364915046365.JavaMail.root@mcs.anl.gov> When we both have time, lets look at the failing run(s) more closely. Im curious whats causing the pattern of failing after 35 app calls. Im also interested in learning how to alleviate any NFS-aggravated shared filesystem problems. - Mike ----- Original Message ----- > From: "Ketan Maheshwari" > To: "Michael Wilde" > Cc: "Swift Devel" > Sent: Tuesday, April 2, 2013 9:43:19 AM > Subject: Re: [Swift-devel] cache already contains error > > > > Hi Mike, > > It works with provider staging. So, it does look like an NFS sync > issue. > > > Thanks, > Ketan > > > > On Mon, Apr 1, 2013 at 10:20 PM, Michael Wilde < wilde at mcs.anl.gov > > wrote: > > > Ketan, > > Can you post a pointer to your code, the Swift log, and the Swift > stdout/err? > > How are you mapping the file "...outdir/out_l0000_0000.0010.out"? > > Does 35 correspond to any of the array bounds? > > Does it fail if you use only one host? (I.e. my first thought was > some kind of NSF sync error). > > Could you try it using local disk and provider staging with the N > hosts? > > > - Mike > > > ----- Original Message ----- > > From: "Ketan Maheshwari" < ketancmaheshwari at gmail.com > > > > > To: "Michael Wilde" < wilde at mcs.anl.gov > > > Cc: "Swift Devel" < swift-devel at ci.uchicago.edu > > > Sent: Monday, April 1, 2013 9:02:24 PM > > Subject: Re: [Swift-devel] cache already contains error > > > > > > Thanks Mike, that fixed the cache issue. However, now I am seeing > > an > > unusual behavior from my Swift run: > > > > > > The ampl run crashes after completing a fixed number of jobs (35 to > > be precise). > > > > > > Some diagnostics: > > > > > > -- It runs to completion when I do a Swift resume. Once again only > > the next 35 jobs complete successfully. On a next resume the rest > > of > > them complete. > > > > > > -- Runs outside of Swift with a bash for-loop using the same > > parameters as in Swift script. > > > > > > -- A catsn script of similar parameters runs to completion without > > any failures. So, nothing seem to be wrong with the OS parameters. > > > > > > I am using a single MCS workstation, no provider staging, no > > coasters. > > > > > > The error message is: > > > > > > Caused by: File not found: > > /nfs2/ketan/powergridapps/swiftscripts/swift.work/inference-20130401-2043-78i7o5m3/shared/outdir/out_l0000_0000.0010.out > > > > > > Which is reflected in the logs as well as in the workdir's info > > files. > > > > > > Has anyone seen this kind of behavior? Any remedial suggestions? > > > > > > Thanks, > > Ketan > > > > > > > > On Mon, Apr 1, 2013 at 5:59 PM, Michael Wilde < wilde at mcs.anl.gov > > > wrote: > > > > > > I think you need to make out 2-dimensional. > > > > Your script is going to evaluate "out[j] = cat(data)" for both i=0 > > and i=1. > > > > The second of those evaluations is probably encountering the "cache > > already contains" for j=0. > > > > If it didnt hit that (ie if you used the concurrent mapper) you'd > > likely then get an error that out[0] is already set. > > > > - Mike > > > > > > > > > > ----- Original Message ----- > > > From: "Ketan Maheshwari" < ketancmaheshwari at gmail.com > > > > To: "Swift Devel" < swift-devel at ci.uchicago.edu > > > > Sent: Monday, April 1, 2013 5:49:36 PM > > > Subject: [Swift-devel] cache already contains error > > > > > > > > > > > > > > > Hi, > > > > > > I am running into the "cache already contains" error when using a > > > nested loop with file mappers. Here is a simple reproduction of > > > the > > > issue with a nested loop variant of catsn.swift: > > > > > > > > > > > > type file; > > > app (file o) cat (file i){ > > > cat @i stdout=@o; > > > } > > > > > > > > > #file out[]; > > > #file out[] > > prefix="f.",suffix=".out">; > > > file out[] > > prefix="f.",suffix=".out">; > > > > > > > > > foreach i in [0:1] { > > > foreach j in [0:1]{ > > > file data<"data.txt">; > > > out[j] = cat(data); > > > } > > > } > > > > > > > > > It runs into the cache error after completing few tasks > > > successfully: > > > > > > $ swift catsn.swift > > > Swift trunk swift-r6410 cog-r3648 > > > > > > > > > RunID: 20130401-1745-7khkyrqc > > > Progress: time: Mon, 01 Apr 2013 17:45:59 -0500 > > > Progress: time: Mon, 01 Apr 2013 17:46:00 -0500 Selecting site:1 > > > Active:1 Finished successfully:2 > > > Execution failed: > > > Exception in cat: > > > Arguments: [data.txt] > > > Host: localhost > > > Directory: catsn-20130401-1745-7khkyrqc/jobs/y/cat-yzf9fg7l > > > Caused by: > > > The cache already contains > > > localhost:catsn-20130401-1745-7khkyrqc/shared/outdir/f.0000.out. > > > cat, catsn.swift, line 14 > > > > > > > > > The cause, I think is that the nested loop triggers the same > > > series > > > of random sequences in mappers code which collides. Both, the > > > simple > > > and the concurrent mappers fail with same message. > > > > > > > > > Does anyone know of a workaround? > > > > > > > > > Thanks, > > > -- > > > Ketan > > > > > > > > > _______________________________________________ > > > Swift-devel mailing list > > > Swift-devel at ci.uchicago.edu > > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > > > > > > > > > > > -- > > Ketan > > > > > > > > > -- > Ketan > > From hategan at mcs.anl.gov Wed Apr 3 14:45:03 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Wed, 03 Apr 2013 12:45:03 -0700 Subject: [Swift-devel] svn issues Message-ID: <1365018303.21131.3.camel@echo> I'm getting Updating '.': svn: E175013: Unable to connect to a repository at URL 'https://svn.ci.uchicago.edu/svn/vdl2/trunk/libexec' svn: E175013: Access to 'https://svn.ci.uchicago.edu/svn/vdl2/trunk/libexec' forbidden Is anybody else seeing this? From wilde at mcs.anl.gov Wed Apr 3 14:47:32 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Wed, 3 Apr 2013 14:47:32 -0500 (CDT) Subject: [Swift-devel] svn issues In-Reply-To: <1365018303.21131.3.camel@echo> Message-ID: <1895432932.4007078.1365018452935.JavaMail.root@mcs.anl.gov> I broke the report for a short while with a bad auth update. Should be fixed now - and I see an update form you just came across so I assume its OK for you now? - Mike ----- Original Message ----- > From: "Mihael Hategan" > To: "Swift Devel" > Sent: Wednesday, April 3, 2013 2:45:03 PM > Subject: [Swift-devel] svn issues > > I'm getting > > Updating '.': > svn: E175013: Unable to connect to a repository at URL > 'https://svn.ci.uchicago.edu/svn/vdl2/trunk/libexec' > svn: E175013: Access to > 'https://svn.ci.uchicago.edu/svn/vdl2/trunk/libexec' forbidden > > Is anybody else seeing this? > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > From hategan at mcs.anl.gov Wed Apr 3 14:50:37 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Wed, 03 Apr 2013 12:50:37 -0700 Subject: [Swift-devel] svn issues In-Reply-To: <1895432932.4007078.1365018452935.JavaMail.root@mcs.anl.gov> References: <1895432932.4007078.1365018452935.JavaMail.root@mcs.anl.gov> Message-ID: <1365018637.21131.4.camel@echo> On Wed, 2013-04-03 at 14:47 -0500, Michael Wilde wrote: > I broke the report for a short while with a bad auth update. > > Should be fixed now - and I see an update form you just came across so I assume its OK for you now? Yes. Sorry, I was going to reply to my email, but was waiting for it to actually get to my inbox (2 min delay on my fetchmail). Mihael From ketancmaheshwari at gmail.com Wed Apr 3 16:20:33 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Wed, 3 Apr 2013 16:20:33 -0500 Subject: [Swift-devel] Swift arrays Message-ID: Trying to get some arithmetic done over arrays. I was testing a simple Swift script as follows: int A[] = [1,1,1]; int B[] = [2,2,2]; foreach b,i in B{ tracef("A[%i] is %i.\n",i, A[i]); tracef("B[%i] is %i.\n",i, B[i]); } Progress: time: Wed, 03 Apr 2013 17:18:13 -0400 A[2] is 1. A[0] is 1. A[1] is 1. Execution failed: org.griphyn.vdl.mapping.InvalidPathException: Invalid path ([1]) for B[1]:int = 2 - Closed swiftscript:tracef, test.swift, line 6 Wondering if this should fail and why. Thanks, -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From wilde at mcs.anl.gov Wed Apr 3 18:04:09 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Wed, 3 Apr 2013 18:04:09 -0500 (CDT) Subject: [Swift-devel] Swift arrays In-Reply-To: Message-ID: <1330164993.4063520.1365030249361.JavaMail.root@mcs.anl.gov> Ketan will explain more, but the problem here seems to be one of case-insensitive names. In: foreach b,i in B{ b and B seem to be referring to the same variable, hence the Invalid Path message. I vaguely recall an older discussion about variable name case sensitivity. What was the latest implementation decision on that? For example, if I declare a and A as variables, I get an "already defined" error, but in the case of the implicit variable declaration of the foreach loop vars, this seems to work differently. - Mike ----- Original Message ----- > From: "Ketan Maheshwari" > To: "Swift Devel" > Sent: Wednesday, April 3, 2013 4:20:33 PM > Subject: [Swift-devel] Swift arrays > > > > Trying to get some arithmetic done over arrays. I was testing a > simple Swift script as follows: > > > > int A[] = [1,1,1]; > int B[] = [2,2,2]; > > > foreach b,i in B{ > tracef("A[%i] is %i.\n",i, A[i]); > tracef("B[%i] is %i.\n",i, B[i]); > } > > > > Progress: time: Wed, 03 Apr 2013 17:18:13 -0400 > A[2] is 1. > A[0] is 1. > A[1] is 1. > Execution failed: > org.griphyn.vdl.mapping.InvalidPathException: Invalid path ([1]) for > B[1]:int = 2 - Closed > swiftscript:tracef, test.swift, line 6 > > > Wondering if this should fail and why. > > > Thanks, > -- > Ketan > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > From ketancmaheshwari at gmail.com Wed Apr 3 21:06:43 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Wed, 3 Apr 2013 21:06:43 -0500 Subject: [Swift-devel] readData from file array Message-ID: Hi, Trying to use the readData method on array of files into variables seems to be failing. Following is an example reproducing the failure: $ cat test2.swift type file; app (file _out) genrand(){ rand stdout=@_out; } file out[] ; int val[]; foreach i in [0:9]{ out[i]=genrand(); val[i]=readData(out[i]); } App rand is mapped to a simple random number generator: $ cat rand.sh #!/bin/bash echo $RANDOM The error message I get on running this is: Could not start execution Failed to convert .swiftx to .kml for test2.swift no such attribute: partialClose in template context [callInternal] I tested readData on scalars which is working ok. Also, if I remove the readData line, the script behaves ok. Thanks, -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From hategan at mcs.anl.gov Wed Apr 3 21:57:39 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Wed, 03 Apr 2013 19:57:39 -0700 Subject: [Swift-devel] Swift arrays In-Reply-To: <1330164993.4063520.1365030249361.JavaMail.root@mcs.anl.gov> References: <1330164993.4063520.1365030249361.JavaMail.root@mcs.anl.gov> Message-ID: <1365044259.3600.3.camel@echo> Seems that you are correct (despite the error message printing the correct type for B): if I say foreach c,..., it finishes fine. I'm wary of fixing this in 0.94, so that leaves trunk. But then if we merge the faster branch to trunk, that problem won't be there any more since things are case-sensitive there. Let me know what you think. Mihael On Wed, 2013-04-03 at 18:04 -0500, Michael Wilde wrote: > Ketan will explain more, but the problem here seems to be one of case-insensitive names. > > In: > > foreach b,i in B{ > > b and B seem to be referring to the same variable, hence the Invalid Path message. > > I vaguely recall an older discussion about variable name case sensitivity. What was the latest implementation decision on that? > > For example, if I declare a and A as variables, I get an "already defined" error, but in the case of the implicit variable declaration of the foreach loop vars, this seems to work differently. > > - Mike > > > ----- Original Message ----- > > From: "Ketan Maheshwari" > > To: "Swift Devel" > > Sent: Wednesday, April 3, 2013 4:20:33 PM > > Subject: [Swift-devel] Swift arrays > > > > > > > > Trying to get some arithmetic done over arrays. I was testing a > > simple Swift script as follows: > > > > > > > > int A[] = [1,1,1]; > > int B[] = [2,2,2]; > > > > > > foreach b,i in B{ > > tracef("A[%i] is %i.\n",i, A[i]); > > tracef("B[%i] is %i.\n",i, B[i]); > > } > > > > > > > > Progress: time: Wed, 03 Apr 2013 17:18:13 -0400 > > A[2] is 1. > > A[0] is 1. > > A[1] is 1. > > Execution failed: > > org.griphyn.vdl.mapping.InvalidPathException: Invalid path ([1]) for > > B[1]:int = 2 - Closed > > swiftscript:tracef, test.swift, line 6 > > > > > > Wondering if this should fail and why. > > > > > > Thanks, > > -- > > Ketan > > > > > > _______________________________________________ > > Swift-devel mailing list > > Swift-devel at ci.uchicago.edu > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel From wilde at mcs.anl.gov Wed Apr 3 22:02:47 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Wed, 3 Apr 2013 22:02:47 -0500 (CDT) Subject: [Swift-devel] Swift arrays In-Reply-To: <1365044259.3600.3.camel@echo> Message-ID: <1996854003.4114687.1365044567456.JavaMail.root@mcs.anl.gov> No problem not fixing it in 0.94. We should document the state of case-sensitivity though. And I hope you can merge the faster branch into trunk as soon as it passes a reasonable test battery. And I'm in favor of "normal" case sensitivity ala C and Java. So if faster-branch is already correctly case-sensitive, I'd say there is nothing to fix here. - Mike ----- Original Message ----- > From: "Mihael Hategan" > To: "Michael Wilde" > Cc: "Ketan Maheshwari" , "Swift Devel" > Sent: Wednesday, April 3, 2013 9:57:39 PM > Subject: Re: [Swift-devel] Swift arrays > > Seems that you are correct (despite the error message printing the > correct type for B): if I say foreach c,..., it finishes fine. > > I'm wary of fixing this in 0.94, so that leaves trunk. But then if we > merge the faster branch to trunk, that problem won't be there any > more > since things are case-sensitive there. > > Let me know what you think. > > Mihael > > On Wed, 2013-04-03 at 18:04 -0500, Michael Wilde wrote: > > Ketan will explain more, but the problem here seems to be one of > > case-insensitive names. > > > > In: > > > > foreach b,i in B{ > > > > b and B seem to be referring to the same variable, hence the > > Invalid Path message. > > > > I vaguely recall an older discussion about variable name case > > sensitivity. What was the latest implementation decision on that? > > > > For example, if I declare a and A as variables, I get an "already > > defined" error, but in the case of the implicit variable > > declaration of the foreach loop vars, this seems to work > > differently. > > > > - Mike > > > > > > ----- Original Message ----- > > > From: "Ketan Maheshwari" > > > To: "Swift Devel" > > > Sent: Wednesday, April 3, 2013 4:20:33 PM > > > Subject: [Swift-devel] Swift arrays > > > > > > > > > > > > Trying to get some arithmetic done over arrays. I was testing a > > > simple Swift script as follows: > > > > > > > > > > > > int A[] = [1,1,1]; > > > int B[] = [2,2,2]; > > > > > > > > > foreach b,i in B{ > > > tracef("A[%i] is %i.\n",i, A[i]); > > > tracef("B[%i] is %i.\n",i, B[i]); > > > } > > > > > > > > > > > > Progress: time: Wed, 03 Apr 2013 17:18:13 -0400 > > > A[2] is 1. > > > A[0] is 1. > > > A[1] is 1. > > > Execution failed: > > > org.griphyn.vdl.mapping.InvalidPathException: Invalid path ([1]) > > > for > > > B[1]:int = 2 - Closed > > > swiftscript:tracef, test.swift, line 6 > > > > > > > > > Wondering if this should fail and why. > > > > > > > > > Thanks, > > > -- > > > Ketan > > > > > > > > > _______________________________________________ > > > Swift-devel mailing list > > > Swift-devel at ci.uchicago.edu > > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > > > _______________________________________________ > > Swift-devel mailing list > > Swift-devel at ci.uchicago.edu > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > From hategan at mcs.anl.gov Wed Apr 3 22:09:44 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Wed, 03 Apr 2013 20:09:44 -0700 Subject: [Swift-devel] Swift arrays In-Reply-To: <1996854003.4114687.1365044567456.JavaMail.root@mcs.anl.gov> References: <1996854003.4114687.1365044567456.JavaMail.root@mcs.anl.gov> Message-ID: <1365044984.4092.1.camel@echo> On Wed, 2013-04-03 at 22:02 -0500, Michael Wilde wrote: > No problem not fixing it in 0.94. We should document the state of case-sensitivity though. Yep. > > And I hope you can merge the faster branch into trunk as soon as it passes a reasonable test battery. Agreed. Let's be specific about what a reasonable test battery is though. From wilde at mcs.anl.gov Wed Apr 3 22:21:40 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Wed, 3 Apr 2013 22:21:40 -0500 (CDT) Subject: [Swift-devel] Test battery for Faster-branch In-Reply-To: <1365044984.4092.1.camel@echo> Message-ID: <214060661.4115418.1365045700158.JavaMail.root@mcs.anl.gov> was: Re: [Swift-devel] Swift arrays > Agreed. Let's be specific about what a reasonable test battery is > though. Indeed. Can you work with Yadu to define this. He has additional stress tests to commit that he's developed over the past month. I suspect much more is needed, but all reasonable additions are a step in the right direction. We can start discussing it in this thread. Yadu started a test plan as a google doc that Ive just shared for editing via the Swift-Devel google group. That doc should reflect what's decided on via this thread. https://docs.google.com/document/d/1wx1QlM8QgrOZI80FrbDSJcrioWtNr8KwZYr7mW24gR0/edit?usp=sharing - Mike From lpesce at uchicago.edu Thu Apr 4 10:47:02 2013 From: lpesce at uchicago.edu (Lorenzo Pesce) Date: Thu, 4 Apr 2013 10:47:02 -0500 Subject: [Swift-devel] Swift arrays In-Reply-To: <1996854003.4114687.1365044567456.JavaMail.root@mcs.anl.gov> References: <1996854003.4114687.1365044567456.JavaMail.root@mcs.anl.gov> Message-ID: <887E124C-C54B-41DF-96BB-8B3BB2721396@uchicago.edu> Let me know when you want to restart trying with the faster branch. 0.94 so far has done well on Beagle (including arrays, readData, with one exception which I think was probably more a case of a waste basket error message problem) On Apr 3, 2013, at 10:02 PM, Michael Wilde wrote: > > No problem not fixing it in 0.94. We should document the state of case-sensitivity though. > > And I hope you can merge the faster branch into trunk as soon as it passes a reasonable test battery. > > And I'm in favor of "normal" case sensitivity ala C and Java. So if faster-branch is already correctly case-sensitive, I'd say there is nothing to fix here. > > - Mike > > > ----- Original Message ----- >> From: "Mihael Hategan" >> To: "Michael Wilde" >> Cc: "Ketan Maheshwari" , "Swift Devel" >> Sent: Wednesday, April 3, 2013 9:57:39 PM >> Subject: Re: [Swift-devel] Swift arrays >> >> Seems that you are correct (despite the error message printing the >> correct type for B): if I say foreach c,..., it finishes fine. >> >> I'm wary of fixing this in 0.94, so that leaves trunk. But then if we >> merge the faster branch to trunk, that problem won't be there any >> more >> since things are case-sensitive there. >> >> Let me know what you think. >> >> Mihael >> >> On Wed, 2013-04-03 at 18:04 -0500, Michael Wilde wrote: >>> Ketan will explain more, but the problem here seems to be one of >>> case-insensitive names. >>> >>> In: >>> >>> foreach b,i in B{ >>> >>> b and B seem to be referring to the same variable, hence the >>> Invalid Path message. >>> >>> I vaguely recall an older discussion about variable name case >>> sensitivity. What was the latest implementation decision on that? >>> >>> For example, if I declare a and A as variables, I get an "already >>> defined" error, but in the case of the implicit variable >>> declaration of the foreach loop vars, this seems to work >>> differently. >>> >>> - Mike >>> >>> >>> ----- Original Message ----- >>>> From: "Ketan Maheshwari" >>>> To: "Swift Devel" >>>> Sent: Wednesday, April 3, 2013 4:20:33 PM >>>> Subject: [Swift-devel] Swift arrays >>>> >>>> >>>> >>>> Trying to get some arithmetic done over arrays. I was testing a >>>> simple Swift script as follows: >>>> >>>> >>>> >>>> int A[] = [1,1,1]; >>>> int B[] = [2,2,2]; >>>> >>>> >>>> foreach b,i in B{ >>>> tracef("A[%i] is %i.\n",i, A[i]); >>>> tracef("B[%i] is %i.\n",i, B[i]); >>>> } >>>> >>>> >>>> >>>> Progress: time: Wed, 03 Apr 2013 17:18:13 -0400 >>>> A[2] is 1. >>>> A[0] is 1. >>>> A[1] is 1. >>>> Execution failed: >>>> org.griphyn.vdl.mapping.InvalidPathException: Invalid path ([1]) >>>> for >>>> B[1]:int = 2 - Closed >>>> swiftscript:tracef, test.swift, line 6 >>>> >>>> >>>> Wondering if this should fail and why. >>>> >>>> >>>> Thanks, >>>> -- >>>> Ketan >>>> >>>> >>>> _______________________________________________ >>>> Swift-devel mailing list >>>> Swift-devel at ci.uchicago.edu >>>> https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel >>>> >>> _______________________________________________ >>> Swift-devel mailing list >>> Swift-devel at ci.uchicago.edu >>> https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel >> >> >> > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel From iraicu at cs.iit.edu Thu Apr 4 15:13:47 2013 From: iraicu at cs.iit.edu (Ioan Raicu) Date: Thu, 04 Apr 2013 15:13:47 -0500 Subject: [Swift-devel] Call for Presentations and Posters: 2nd Greater Chicago Area System Research Workshop (GCASR) 2013 Message-ID: <515DDEFB.906@cs.iit.edu> ** *Call for Presentations and Posters* *2nd Greater Chicago Area System Research Workshop (GCASR) 2013* *Northwestern University, Evanston IL -- May 3rd, 2013 * *http://datasys.cs.iit.edu/events/GCASR13/index.html * ** *Overview* It is our pleasure to invite you to join us for the 2nd Greater Chicago Area Systems Research (GCASR) Workshop. This workshop is a great opportunity for leading-edge systems researchers in Chicago to get acquainted, share results, and find synergies. We plan to have a renowned researcher as the keynote, as well as three presentation sessions, one Ph.D student poster session, and a panel discussion. For last year's workshop agenda, please click here . *Topics* This workshop will focus on Computer Systems, broadly construed, and related areas. Topics of interest include, but are not limited to: * Operating Systems * Architecture * Networking * Distributed Systems * Grids/Clouds/Supercomputing (including HPC/MTC/HTC) * Reconfigurable and Embedded Systems * Programming Languages * Storage Systems * Databases *Important Dates* * *Abstract submission: *April 11th, 2013 * *Acceptance notification: *April 15th, 2013 * *Workshop date: *May 3rd, 2013 *Location* The 2nd Greater Chicago Area Systems Research (GCASR) Workshop will take place at: *Cohen Commons, Technological Institute L482* *Northwestern University* *2145 Sheridan Road* *Evanston, IL 60208* Northwestern University is located on the North Shore of the Chicago area. The Technological Institute can be conveniently accessed via the L Train (take the Purple Line to Noyes Stop, walk East 4 blocks to the Technological Institute) or by car or Taxi. Detailed directions and parking information can be found at http://www.eecs.northwestern.edu/resources/for-visitors.html. *Abstract Submission* Authors are invited to submit abstracts (for either oral or poster presentation) of not more than 300 words, plus a biography of the presentation author. The abstract must be submitted online through the Google Form at https://docs.google.com/forms/d/13tE5lHtxjR-vgzNK5YzD8hON8F7uHbL_9laAvWDfp-M/viewform before the deadline of April 11th, 2013 at 11:59PM CDT. Abstracts will be reviewed for relevance and institutional coverage. Notifications of the decisions will be sent out by April 15th, 2013. For more information, please see http://datasys.cs.iit.edu/events/GCASR13/. *Organization * *Chairs* * Ioan Raicu , Illinois Institute of Technology & Argonne National Laboratory, USA * Nikos Hardavellas , Northwestern University, USA *Steering Committee* * Fabi?n E. Bustamante , Northwestern University, USA * Andrew A. Chien , University of Chicago & Argonne National Laboratory, USA * Peter Dinda , Northwestern University, USA * Jakob Eriksson , University of Illinois at Chicago, USA * Ian Foster , University of Chicago & Argonne National Laboratory, USA * Nikos Hardavellas , Northwestern University, USA * Ioan Raicu , Illinois Institute of Technology & Argonne National Laboratory, USA * Xian-He Sun , Illinois Institute of Technology, USA *Sponsors* The GCASR 2013 workshop has received generous financial support from Robert R. McCormick School of Engineering and Applied Science at Northwestern University, and the Department of Electrical Engineering and Computer Science at Northwestern University. -- ================================================================= Ioan Raicu, Ph.D. Assistant Professor, Illinois Institute of Technology (IIT) Guest Research Faculty, Argonne National Laboratory (ANL) ================================================================= Data-Intensive Distributed Systems Laboratory, CS/IIT Distributed Systems Laboratory, MCS/ANL ================================================================= Editor: IEEE TCC, Springer JoCCASA Chair: IEEE/ACM MTAGS, ACM ScienceCloud, IEEE/ACM DataCloud ================================================================= Cel: 1-847-722-0876 Office: 1-312-567-5704 Email: iraicu at cs.iit.edu Web: http://www.cs.iit.edu/~iraicu/ Web: http://datasys.cs.iit.edu/ LinkedIn: http://www.linkedin.com/in/ioanraicu Google: http://scholar.google.com/citations?user=jE73HYAAAAAJ ================================================================= ================================================================= -------------- next part -------------- An HTML attachment was scrubbed... URL: From davidk at ci.uchicago.edu Thu Apr 4 17:41:16 2013 From: davidk at ci.uchicago.edu (David Kelly) Date: Thu, 4 Apr 2013 17:41:16 -0500 (CDT) Subject: [Swift-devel] Formatting floats In-Reply-To: <18361937.4497417.1365114912707.JavaMail.root@ci.uchicago.edu> Message-ID: <1199631010.4499154.1365115276436.JavaMail.root@ci.uchicago.edu> Hello, Is it possible to format floats in a printf() like way in Swift? I tried using @sprintf to print to two decimal places, but it doesn't seem to do what I want: float f=0.0; string s = @sprintf("%.2f", f); Execution failed: tracef(): Unknown format: %. swiftscript:sprintf, test.swift, line 2 assignment, test.swift, line 3 David -------------- next part -------------- An HTML attachment was scrubbed... URL: From wozniak at mcs.anl.gov Thu Apr 4 17:48:54 2013 From: wozniak at mcs.anl.gov (Justin M Wozniak) Date: Thu, 04 Apr 2013 17:48:54 -0500 Subject: [Swift-devel] Formatting floats In-Reply-To: <1199631010.4499154.1365115276436.JavaMail.root@ci.uchicago.edu> References: <1199631010.4499154.1365115276436.JavaMail.root@ci.uchicago.edu> Message-ID: <515E0356.9090508@mcs.anl.gov> The current function does not do this. My thought was it is good to be able to represent Swift-specific types. But it would also be good to have a function that simply uses Java's String.format(). On 04/04/2013 05:41 PM, David Kelly wrote: > Hello, > > Is it possible to format floats in a printf() like way in Swift? I > tried using @sprintf to print to two decimal places, but it doesn't > seem to do what I want: > > float f=0.0; > string s = @sprintf("%.2f", f); > > Execution failed: > tracef(): Unknown format: %. > swiftscript:sprintf, test.swift, line 2 > assignment, test.swift, line 3 > > David > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel -- Justin M Wozniak -------------- next part -------------- An HTML attachment was scrubbed... URL: From ketancmaheshwari at gmail.com Wed Apr 10 18:47:26 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Wed, 10 Apr 2013 18:47:26 -0500 Subject: [Swift-devel] Intrepid Message-ID: Hi All, Trying to run Swift on Intrepid but jobs are failing. I am using coasters with local:cobalt provider. The last error message I am getting is: Job failed with an exit code of 30 Apart from this, the std.err also shows this error message: Failed to open log file (/home/ketan/lab/catsnsleep/worker-0410-2811160-000000-69be4b00.log): Read-only file system at /home/ketan/.globus/coasters/cscript6918.pl line 416 Line 416 on that perl is: open(LOG, ">>$LOG") or die "Failed to open log file ($LOG): $!"; The logs, sources, configs are here: http://www.mcs.anl.gov/~ketan/intrepidlogs.tgz The sites file is yanked from Jon Monette's homedir on Intrepid since the dir showed signs of working version. It was with Swift 0.93 and my current test is with 0.94 RC though. Thanks much for any help on this. Regards, Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From wozniak at mcs.anl.gov Wed Apr 10 19:09:51 2013 From: wozniak at mcs.anl.gov (Justin M Wozniak) Date: Wed, 10 Apr 2013 19:09:51 -0500 Subject: [Swift-devel] Intrepid In-Reply-To: References: Message-ID: <5165FF4F.4040402@mcs.anl.gov> Can you try using /intrepid-fs0 ? On 04/10/2013 06:47 PM, Ketan Maheshwari wrote: > Hi All, > > Trying to run Swift on Intrepid but jobs are failing. I am using > coasters with local:cobalt provider. > > The last error message I am getting is: > > Job failed with an exit code of 30 > > Apart from this, the std.err also shows this error message: > > Failed to open log file > (/home/ketan/lab/catsnsleep/worker-0410-2811160-000000-69be4b00.log): > Read-only file system at /home/ketan/.globus/coasters/cscript6918.pl > line 416 > > Line 416 on that perl is: > > open(LOG, ">>$LOG") or die "Failed to open log file ($LOG): $!"; > > > The logs, sources, configs are here: > > http://www.mcs.anl.gov/~ketan/intrepidlogs.tgz > > > The sites file is yanked from Jon Monette's homedir on Intrepid since > the dir showed signs of working version. It was with Swift 0.93 and my > current test is with 0.94 RC though. > > > Thanks much for any help on this. > > Regards, > Ketan > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel -- Justin M Wozniak -------------- next part -------------- An HTML attachment was scrubbed... URL: From wilde at mcs.anl.gov Thu Apr 11 11:09:43 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Thu, 11 Apr 2013 11:09:43 -0500 (CDT) Subject: [Swift-devel] Has anyone run LAMMPS under Swift? Message-ID: <1907456632.1416364.1365696583739.JavaMail.root@mcs.anl.gov> Hi All, If youve run LAMMPS under Swift, can you send me some example scripts? Thanks, - Mike -- Michael Wilde Computation Institute, University of Chicago Mathematics and Computer Science Division Argonne National Laboratory From davidk at ci.uchicago.edu Thu Apr 11 11:16:14 2013 From: davidk at ci.uchicago.edu (David Kelly) Date: Thu, 11 Apr 2013 11:16:14 -0500 (CDT) Subject: [Swift-devel] Has anyone run LAMMPS under Swift? In-Reply-To: <1907456632.1416364.1365696583739.JavaMail.root@mcs.anl.gov> Message-ID: <1033322260.1256685.1365696974074.JavaMail.root@ci.uchicago.edu> Mike, The CMTS dimers script in SVN uses LAMPPS. It uses a non-standard version of LAMPPS though, with some custom patches applied. David ----- Original Message ----- > From: "Michael Wilde" > To: "Swift Devel" > Sent: Thursday, April 11, 2013 11:09:43 AM > Subject: [Swift-devel] Has anyone run LAMMPS under Swift? > Hi All, > If youve run LAMMPS under Swift, can you send me some example > scripts? > Thanks, > - Mike > -- > Michael Wilde > Computation Institute, University of Chicago > Mathematics and Computer Science Division > Argonne National Laboratory > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel -------------- next part -------------- An HTML attachment was scrubbed... URL: From ketancmaheshwari at gmail.com Thu Apr 11 11:32:34 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Thu, 11 Apr 2013 11:32:34 -0500 Subject: [Swift-devel] Intrepid In-Reply-To: <5165FF4F.4040402@mcs.anl.gov> References: <5165FF4F.4040402@mcs.anl.gov> Message-ID: Thank you. Moving everything to /intrepid-fs0/users/ketan/scratch worked on Intrepid. Found from an old Beagle discussion, setting SWIFT_USERHOME to /scratch path was also required in such cases where the /home filesystem was not writable by compute nodes. On Wed, Apr 10, 2013 at 7:09 PM, Justin M Wozniak wrote: > > Can you try using /intrepid-fs0 ? > > On 04/10/2013 06:47 PM, Ketan Maheshwari wrote: > > Hi All, > > Trying to run Swift on Intrepid but jobs are failing. I am using coasters > with local:cobalt provider. > > The last error message I am getting is: > > Job failed with an exit code of 30 > > Apart from this, the std.err also shows this error message: > > Failed to open log file > (/home/ketan/lab/catsnsleep/worker-0410-2811160-000000-69be4b00.log): > Read-only file system at /home/ketan/.globus/coasters/cscript6918.pl line > 416 > > Line 416 on that perl is: > > open(LOG, ">>$LOG") or die "Failed to open log file ($LOG): $!"; > > > The logs, sources, configs are here: > > http://www.mcs.anl.gov/~ketan/intrepidlogs.tgz > > The sites file is yanked from Jon Monette's homedir on Intrepid since > the dir showed signs of working version. It was with Swift 0.93 and my > current test is with 0.94 RC though. > > > Thanks much for any help on this. > > Regards, > Ketan > > > _______________________________________________ > Swift-devel mailing listSwift-devel at ci.uchicago.eduhttps://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > > -- > Justin M Wozniak > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From wilde at mcs.anl.gov Sat Apr 13 12:13:51 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Sat, 13 Apr 2013 12:13:51 -0500 (CDT) Subject: [Swift-devel] Need help to enable provenance In-Reply-To: <1668257296.2638748.1365873082227.JavaMail.root@mcs.anl.gov> Message-ID: <1560323762.2638766.1365873231387.JavaMail.root@mcs.anl.gov> Hi All, I cant get provenance enabled in either trunk or the head revision of 0.94. Can anyone help me with this? Ive got the necessary log4j entries set, and have provenence.log=true in my -config file. But my .kml files start off with: I last had this working with this swift: Swift trunk swift-r6171 cog-r3560 (cog modified locally) Two questions: - has anyone specifically disabled provenance? - where is the .kml comment line above coming from? (Im tracking it down now) Can anyone provide some guidance on enabling provenance logging? Specifically, Im looking for lines like this in the .log file with the provenanceid attribute set: catsn-20130129-1033-sgtqveud.log:2013-01-29 10:33:26,978-0600 INFO swift PARAM thread=0-3-6-1 direction=input variable=i provenanceid=dataset:20130129-1033-gxhdq72d:720000000038 Thanks, - Mike -- Michael Wilde Computation Institute, University of Chicago Mathematics and Computer Science Division Argonne National Laboratory From hategan at mcs.anl.gov Sat Apr 13 15:24:35 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Sat, 13 Apr 2013 13:24:35 -0700 Subject: [Swift-devel] Need help to enable provenance In-Reply-To: <1560323762.2638766.1365873231387.JavaMail.root@mcs.anl.gov> References: <1560323762.2638766.1365873231387.JavaMail.root@mcs.anl.gov> Message-ID: <1365884675.11612.2.camel@echo> I did a quick test with 0.94 and if I set provenance.log=true, I get So I'm not quite sure what's happening in your case. I would check simple things like making sure you are editing the right configuration file, that you are not overriding -provenance.log on the command line, etc. Posting the log might help. Mihael On Sat, 2013-04-13 at 12:13 -0500, Michael Wilde wrote: > Hi All, > > I cant get provenance enabled in either trunk or the head revision of 0.94. Can anyone help me with this? > > Ive got the necessary log4j entries set, and have provenence.log=true in my -config file. > > But my .kml files start off with: > > > > I last had this working with this swift: > > Swift trunk swift-r6171 cog-r3560 (cog modified locally) > > Two questions: > > - has anyone specifically disabled provenance? > > - where is the .kml comment line above coming from? (Im tracking it down now) > > Can anyone provide some guidance on enabling provenance logging? > > Specifically, Im looking for lines like this in the .log file with the provenanceid attribute set: > > catsn-20130129-1033-sgtqveud.log:2013-01-29 10:33:26,978-0600 INFO swift PARAM thread=0-3-6-1 direction=input variable=i provenanceid=dataset:20130129-1033-gxhdq72d:720000000038 > > Thanks, > > - Mike > > > From wilde at mcs.anl.gov Sat Apr 13 16:21:13 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Sat, 13 Apr 2013 16:21:13 -0500 (CDT) Subject: [Swift-devel] Need help to enable provenance In-Reply-To: <1365884675.11612.2.camel@echo> Message-ID: <1984312271.2646359.1365888073565.JavaMail.root@mcs.anl.gov> > So I'm not quite sure what's happening in your case. I would check > simple things like making sure you are editing the right > configuration > file, that you are not overriding -provenance.log on the command > line, > etc. Indeed, that seems to have been the problem: I forgot that I had created a $HOME/.swift/swift.properties file earlier in the week. But the resulting properties ordering seems to be less than clear and desirable behavior: even though (a) I was specifying via -config a properties file with provenance=true, (b) the .log file confirmed this, it seems that swift took at least the provenance setting from the $HOME/.swift/swift.properties file. We should examine this in more detail. I would have expected that properties are read in this order: $SWIFT_HOME/etc/swift.properties $HOME/.swift/swift.properties swift -config swift.properties Im not sure, but perhaps the provenanceEnabled state is getting latched in to "false" (from an initial null) in some class(es) when the second properties file above is processed, and the -config file's provenance.value is being ignored? Or maybe its just that the second file above has the final say for all values? Needs more experimentation, but something doesnt seem right and logical here. - Mike From hategan at mcs.anl.gov Sat Apr 13 16:41:50 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Sat, 13 Apr 2013 14:41:50 -0700 Subject: [Swift-devel] Need help to enable provenance In-Reply-To: <1984312271.2646359.1365888073565.JavaMail.root@mcs.anl.gov> References: <1984312271.2646359.1365888073565.JavaMail.root@mcs.anl.gov> Message-ID: <1365889310.12668.1.camel@echo> On Sat, 2013-04-13 at 16:21 -0500, Michael Wilde wrote: > > So I'm not quite sure what's happening in your case. I would check > > simple things like making sure you are editing the right > > configuration > > file, that you are not overriding -provenance.log on the command > > line, > > etc. > > Indeed, that seems to have been the problem: I forgot that I had created a $HOME/.swift/swift.properties file earlier in the week. > > But the resulting properties ordering seems to be less than clear and desirable behavior: even though (a) I was specifying via -config a properties file with provenance=true, (b) the .log file confirmed this, it seems that swift took at least the provenance setting from the $HOME/.swift/swift.properties file. > > We should examine this in more detail. I would have expected that properties are read in this order: > > $SWIFT_HOME/etc/swift.properties > $HOME/.swift/swift.properties > swift -config swift.properties Agreed. That should be the order, and the absence of a property in a higher priority configuration file should override existing properties in other files. > > Im not sure, but perhaps the provenanceEnabled state is getting latched in to "false" (from an initial null) in some class(es) when the second properties file above is processed, and the -config file's provenance.value is being ignored? Or maybe its just that the second file above has the final say for all values? Needs more experimentation, but something doesnt seem right and logical here. > > - Mike > From wilde at mcs.anl.gov Wed Apr 17 15:51:31 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Wed, 17 Apr 2013 15:51:31 -0500 (CDT) Subject: [Swift-devel] auto-coaster bootstrap for stampede cluster In-Reply-To: <1390216294.3338592.1366231530112.JavaMail.root@ci.uchicago.edu> Message-ID: <1617183787.3432082.1366231891105.JavaMail.root@mcs.anl.gov> I'm moving this topic to swift-devel, so others, in particular Mihael, can weigh in. - Mike ----- Forwarded Message ----- From: "David Kelly" To: "Ketan Maheshwari" Cc: "Wilde" Sent: Wednesday, April 17, 2013 3:45:30 PM Subject: Fwd: auto-coaster bootstrap for stamped Hey Ketan, Mike mentioned that you were interested in running remotely to Stampede via ssh-cl. Normally we could use ssh-cl like any other site, but the problem we ran into here is that we can't run Swift on the stampede head node. We need to ssh-cl AND also start swift on a remote worker node, which is a setup that hasn't been tested very much. I believe you've used start-coaster-service before when we were running on ec2. You can this configuration for Stampede too. Modify coaster-service.conf to set WORKER_NODE=slurm, WORKER_RELAY_HOST=stampede.tacc.utexas.edu, and it will generate a slurm script, scp it to stampede, and remotely start swift on a worker node. I'll see if I can find an example config file for this. With automatic coaters it's a bit more complicated and completely untested as far as I know. You may be able to use gram2. This worked on Ranger, but haven't tried yet on Stampede. Mike mentioned in the email below you may be able to change the ssh-cl provider to add some kind of prefix command (srun). Maybe you can modify your PATH so the 'ssh' command is actually a wrapper you created and does something sneaky. You may also be able to add a prefix command to cog/modules/provider-coaster/resources/bootstrap.sh. Hopefully this can help you get started - let me know if any of this works for you, curious to see how we can get it working well. David ----- Forwarded Message ----- From: "Michael Wilde" To: "David Kelly" Sent: Tuesday, April 16, 2013 10:59:22 AM Subject: auto-coaster bootstrap for stamped was: Re: Another item for the to-do list David, thanks for the details. Im wondering, for systems like stampede, could automatic coasters work to it (eg from swift.rcc) by adding a sinteractive or srun command into the middle of the ssh command generated by the ssh-cl parameter? ie instead of doing ssh -sshargsgere auto-boostrap-coaster-stuff-here.sh do: ssh -sshargsgere srun auto-boostrap-coaster-stuff-here.sh ? > This is the only mode that I've been able to test on Stampede so far. > I will experiment more the others when Stampede is back up. Others meaning GRAM? Perhaps using myproxy-logon? That *should* work out of the box but we've not tested GRAM in ages so it probably doesnt. Lets keep this lower on the prio list. I just want to be sure we have a ticket for this. Please create one if not - thanks. From iraicu at cs.iit.edu Fri Apr 19 12:36:14 2013 From: iraicu at cs.iit.edu (Ioan Raicu) Date: Fri, 19 Apr 2013 12:36:14 -0500 Subject: [Swift-devel] Call for Participation: 2nd Greater Chicago Area System Research Workshop (GCASR) 2013 Message-ID: <5171808E.8050105@cs.iit.edu> 2nd Greater Chicago Area System Research Workshop (GCASR) 2013 *At Northwestern University, Evanston IL -- May 3rd, 2013 *http://datasys.cs.iit.edu/events/GCASR13/ Overview It is our pleasure to invite you to join us for the 2nd Greater Chicago Area Systems Research (GCASR) Workshop. This workshop is a great opportunity for leading-edge systems researchers in Chicago to get acquainted, share results, and find synergies. The workshop program has been posted. The workshop is free of charge (courtesy of our generous sponsors from Northwestern University <#Sponsors>); please register (https://docs.google.com/forms/d/1Q6cl0NlgSz_N-mwY6ccrM7h0-J_xhnDuc-of7-5lLYs/viewform) by April 25th, 2013 if you would like to attend. Program We are pleased to announce that the final program (http://datasys.cs.iit.edu/events/GCASR13/program.html) is composed of 1 keynote presentation , 15 oral presentations <#Oral_Presentations>, and 39 poster presentations (separated into session 1 <#Poster_Session_1> and session 2 <#Poster_Session_2>). We have 57 workshop presenters, representing 10 different institutions (ANL, DePaul, Google, IIT, Northwestern, Notre Dame, Purdue, UChicago, UIC, and UIUC). The workshop will start at 8AM with breakfast, and will end at 6:15PM. Keynote Presentation * Pete Beckman , Argonne National Laboratory Oral Presentations Time Description 8:00 - 8:30 Breakfast 8:30 - 8:40 Opening Remarks Nikos Hardavellas, Northwestern Ioan Raicu, IIT 8:40 - 9:25 Keynote Pete Beckman, ANL Session 1: Architecture, Embedded and Reconfigurable Systems 9:25 - 9:50 *The 10x10 Project: Can we have both extreme energy efficiency and programmability? *Andrew A. Chien, UChicago 9:50 - 10:15 *Collectively Achieve Fairest Distribution of Discrete Resources with Limited Sharing: Complexity and Protocols* Wenjing Rao, UIC 10:15 - 10:30 Break Session 2: HPC, Cloud, and Warehouse-scale Computing 10:30 - 10:55 *Warehouse-scale Computing: Challenges and Opportunities* Nicholas Kidd, Google 10:55 - 11:20 *Scaling Up Without Blowing Up* Douglas Thain, Notre Dame 11:20 - 11:45 *The Case for Limping-Hardware Tolerant Clouds * Haryadi Gunawi, UChicago 11:45 - 12:10 *Toward Smart HPC via Active Learning and Intelligent Scheduling* Zhiling Lan, IIT 12:10 - 12:55 Lunch 12:55 - 1:40 Poster Session 1 Distributed Systems and Storage <#Poster_Session_1> Session 3: Networking and Storage 1:40 - 2:05 *Pushing experiments to the Internet's edge * Fabian E. Bustamante, Northwestern 2:05 - 2:30 *Increasing Network Resiliency by Optimally Assigning Diverse Variants to Routing Nodes * Cristina Nita-Rotaru, Purdue 2:30 - 2:55 *Parallel Networks and Storage for Predictable End-to-End Data Movement * Eun-Sung Jung, ANL 2:55 - 3:20 *System Design Considerations of Big Data Processing *Xian-He Sun, IIT 3:20 - 4:05 Break Poster Session 2 Architecture, Data-bases, Networking, OS, Programming Languages, and Security <#Poster_Session_2> Session 4: Operating Systems and Security 4:05 - 4:30 *Recent Research in the V3VEE Project *Peter Dinda, Northwestern 4:30 - 4:55 *Ethos: A Layered Approach to Secure Applications *Jon A Solworth, UIC 4:55 - 5:20 *DroidChameleon: Evaluating Android Anti-malware against Transformation Attacks *Yan Chen, Northwestern Session 5: Programming Languages and Database Systems 5:20 - 5:45 *Diderot: a Domain-Specific Language for Portably Parallel Scientific Visualization and Image Analysis *Gordon Kindlmann, UChicago 5:45 - 6:10 *The (Potential) Perks of Integrating Provenance Support into Database Engines *Boris Glavic, IIT 6:10 Closing Remarks Nikos Hardavellas, Northwestern Ioan Raicu, IIT Poster Session 1 Distributed Systems and Storage Title Name Institution A Decoupled Execution Paradigm for Data-Intensive High-End Computing Yanlong Yin IIT Distributed Kev-Value Store on Cloud Tonglin Li IIT Enabling and Optimizing Parallel File Write in HDFS Xi Yang IIT FusionFS: a distributed file system for large scale data-intensive computing Dongfang Zhao IIT HPC Analytics for Extreme Scale Computing Li Yu IIT MATRIX: MAny-Task computing execution fabRIc at eXascales Anupam Rajendran IIT Measuring Power Consumption on IBM Blue Gene/Q Sean Wallace IIT NoVoHT: a Lightweight Dynamic Persistent NoSQL Key/Value Store *Kevin Brandstatter IIT Power-Aware Job Scheduling on Production HPC Systems Zhou Zhou IIT Toward Petascale Cosmology Simulations Jingjin Wu IIT Towards a Provenance-aware Distributed Filesystem Chen Shou IIT Towards Distributed Batch Scheduling Xiaobing Zhou IIT Understanding the cost of the cloud for scientific applications Iman Sadooghi IIT Community sensing under (soft) control John Rula Northwestern Blockus: Efficient Out-of-Core Computation Using Blocks Erik Bodzsar UChicago Exploiting Global View for Resilience (GVR) Hajime Fujita UChicago The Global View Resilience Model Zachary Rubenstein UChicago Linking and Loading Scientific Workflows Casey Robinson Notre Dame Scaling Work Queue for the Cloud with Hierarchy Michael Albrecht Notre Dame Software Engineering Challenges and Tools in Distributed Computing Workspace for Civil Engineering Applications Peter Sempolinski Notre Dame Poster Session 2 Architecture, Data-bases, Networking, OS, Programming Languages, and Security Title Name Institution Mastering chaos with cost-effective sampling Mona Rahimi DePaul Accelerating Scientific Workflow Applications with GPUs *Dustin Shahidehpour IIT Enabling Dynamic Memory Management Support for MTC on NVIDIA GPUs *Ben Grimmer IIT GeMTC: GPU enabled Many-Task Computing Scott J. Krieder IIT Understanding the Costs of Many-Task Computing Workloads on Intel Xeon Phi Coprocessors *Jeffrey Johnson IIT Characterizing Broadband Services with Dasu Zachary Bischof Northwestern Elastic Fidelity: Trading-off Computational Accuracy for Energy Reduction Georgios Tziantzioulis Northwestern Experiments at the Internet's Edge with Dasu Mario A. Sanchez Northwestern Galaxy: Pushing the Power and Bandwidth Walls with Optically Connected Disintegrated Processors Yigit Demir Northwestern Shifting GEARS to Enable Guest-context Virtual Services Kyle Hale Northwestern Understanding Timing Errors Patterns Yuanbo Fan Northwestern VMM-based Emulation of Intel Hardware Transactional Memory Maciej Swiech Northwestern Compiler optimization for massively parallel data flow Timothy Armstrong UChicago The 10x10 Project: Can we have both extreme energy efficiency and programmability? Raihan ur Rasool UChicago When is Multi-version Checkpointing Needed? Guoming Lu UChicago MinimaLT: Minimal Latency Networking Through Better Security Xu Zhang UIC Scalable Authentication Infrastructure Wenyuan Fei UIC Simple and Secure Networking Yaohua Li UIC Automatic Memory Deduplication Support in Hypervisor Fangzhou Yao UIUC Note that * denotes undergraduate students. Location The 2nd Greater Chicago Area Systems Research (GCASR) Workshop will take place at: Louis Room, Norris University Center 205B Northwestern University 1999 Campus Drive Evanston, IL 60208 Map to Norris University Center: http://maps.northwestern.edu/#latlngz=42.053%2C-87.673%2C17&lookupid=110 Norris Floor Plan http://www.norris.northwestern.edu/about/floor-plans/ Northwestern University is located on the North Shore of the Chicago area. Northwestern can be conveniently accessed via the L Train (take the Purple Line to Noyes Stop, walk East 4 blocks to the main campus) or by car or Taxi. Detailed directions and parking information can be found at http://www.eecs.northwestern.edu/resources/for-visitors.html. Organization *Chairs* * Ioan Raicu , Illinois Institute of Technology & Argonne National Laboratory, USA * Nikos Hardavellas , Northwestern University, USA *Steering Committee* * Fabi?n E. Bustamante , Northwestern University, USA * Andrew A. Chien , University of Chicago & Argonne National Laboratory, USA * Peter Dinda , Northwestern University, USA * Jakob Eriksson , University of Illinois at Chicago, USA * Ian Foster , University of Chicago & Argonne National Laboratory, USA * Nikos Hardavellas , Northwestern University, USA * Ioan Raicu , Illinois Institute of Technology & Argonne National Laboratory, USA * Xian-He Sun , Illinois Institute of Technology, USA -- ================================================================= Ioan Raicu, Ph.D. Assistant Professor, Illinois Institute of Technology (IIT) Guest Research Faculty, Argonne National Laboratory (ANL) ================================================================= Data-Intensive Distributed Systems Laboratory, CS/IIT Distributed Systems Laboratory, MCS/ANL ================================================================= Editor: IEEE TCC, Springer JoCCASA Chair: IEEE/ACM MTAGS, ACM ScienceCloud, IEEE/ACM DataCloud ================================================================= Cel: 1-847-722-0876 Office: 1-312-567-5704 Email: iraicu at cs.iit.edu Web: http://www.cs.iit.edu/~iraicu/ Web: http://datasys.cs.iit.edu/ LinkedIn: http://www.linkedin.com/in/ioanraicu Google: http://scholar.google.com/citations?user=jE73HYAAAAAJ ================================================================= ================================================================= -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- An HTML attachment was scrubbed... URL: From ketancmaheshwari at gmail.com Fri Apr 19 20:32:58 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Fri, 19 Apr 2013 20:32:58 -0500 Subject: [Swift-devel] midway jobs not directed to right site Message-ID: Hi, I am trying to run a script where some lightweight jobs were destined to run on localhost while others were to run on sandyb/westmere queues in midway. I do this by putting the right handle before the app name in my tc. The corresponding pool entry being in the sites.xml However, it looks like all the jobs in the script are running on localhost instead. When I comment out the localhost part from the sies.xml and the script, I correctly get them to run on sandyb/westmere queues. In order to verify this, I tried to run a Swift script with three jobs each destined to westmere, sandyb and localhost echoing a message and printing the hostname. I do find the same behavior in this case too. Jobs were all running on localhost and none on the queues. See the above mentioned example bundle with logs and config files here: http://www.mcs.anl.gov/~ketan/sites_exp.tgz Should run straightaway with ./run.midway and . in PATH. Feature or bug? -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From hategan at mcs.anl.gov Sat Apr 20 00:59:43 2013 From: hategan at mcs.anl.gov (Mihael Hategan) Date: Fri, 19 Apr 2013 22:59:43 -0700 Subject: [Swift-devel] midway jobs not directed to right site In-Reply-To: References: Message-ID: <1366437583.28696.3.camel@echo> They probably go to the right site, but all of those sites point to whatever coaster service runs on localhost of which there is only one (once one is started, whenever you try to run a coaster job on localhost, the already started service will be used). So you might want to change the last site to use the local provider instead. Mihael On Fri, 2013-04-19 at 20:32 -0500, Ketan Maheshwari wrote: > Hi, > > I am trying to run a script where some lightweight jobs were destined to > run on localhost while others were to run on sandyb/westmere queues in > midway. > > I do this by putting the right handle before the app name in my tc. The > corresponding pool entry being in the sites.xml > > However, it looks like all the jobs in the script are running on localhost > instead. > > When I comment out the localhost part from the sies.xml and the script, I > correctly get them to run on sandyb/westmere queues. > > In order to verify this, I tried to run a Swift script with three jobs each > destined to westmere, sandyb and localhost echoing a message and printing > the hostname. I do find the same behavior in this case too. Jobs were all > running on localhost and none on the queues. > > See the above mentioned example bundle with logs and config files here: > > http://www.mcs.anl.gov/~ketan/sites_exp.tgz > > Should run straightaway with ./run.midway and . in PATH. > > > Feature or bug? > Hi, > > > I am trying to run a script where some lightweight jobs were destined > to run on localhost while others were to run on sandyb/westmere queues > in midway. > > > I do this by putting the right handle before the app name in my tc. > The corresponding pool entry being in the sites.xml > > > However, it looks like all the jobs in the script are running on > localhost instead. > > > When I comment out the localhost part from the sies.xml and the > script, I correctly get them to run on sandyb/westmere queues. > > > In order to verify this, I tried to run a Swift script with three jobs > each destined to westmere, sandyb and localhost echoing a message and > printing the hostname. I do find the same behavior in this case too. > Jobs were all running on localhost and none on the queues. > > > See the above mentioned example bundle with logs and config files > here: > > > http://www.mcs.anl.gov/~ketan/sites_exp.tgz > > > > Should run straightaway with ./run.midway and . in PATH. > > > > > Feature or bug? > -- > Ketan > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel From ketancmaheshwari at gmail.com Sat Apr 20 10:48:33 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Sat, 20 Apr 2013 10:48:33 -0500 Subject: [Swift-devel] midway jobs not directed to right site In-Reply-To: <1366437583.28696.3.camel@echo> References: <1366437583.28696.3.camel@echo> Message-ID: Thanks Mihael! Works after turning off provider staging and setting filesystem provider in each of the queue pools. On Sat, Apr 20, 2013 at 12:59 AM, Mihael Hategan wrote: > They probably go to the right site, but all of those sites point to > whatever coaster service runs on localhost of which there is only one > (once one is started, whenever you try to run a coaster job on > localhost, the already started service will be used). > > So you might want to change the last site to use the local provider > instead. > > Mihael > > On Fri, 2013-04-19 at 20:32 -0500, Ketan Maheshwari wrote: > > Hi, > > > > I am trying to run a script where some lightweight jobs were destined to > > run on localhost while others were to run on sandyb/westmere queues in > > midway. > > > > I do this by putting the right handle before the app name in my tc. The > > corresponding pool entry being in the sites.xml > > > > However, it looks like all the jobs in the script are running on > localhost > > instead. > > > > When I comment out the localhost part from the sies.xml and the script, I > > correctly get them to run on sandyb/westmere queues. > > > > In order to verify this, I tried to run a Swift script with three jobs > each > > destined to westmere, sandyb and localhost echoing a message and printing > > the hostname. I do find the same behavior in this case too. Jobs were all > > running on localhost and none on the queues. > > > > See the above mentioned example bundle with logs and config files here: > > > > http://www.mcs.anl.gov/~ketan/sites_exp.tgz > > > > Should run straightaway with ./run.midway and . in PATH. > > > > > > Feature or bug? > > Hi, > > > > > > I am trying to run a script where some lightweight jobs were destined > > to run on localhost while others were to run on sandyb/westmere queues > > in midway. > > > > > > I do this by putting the right handle before the app name in my tc. > > The corresponding pool entry being in the sites.xml > > > > > > However, it looks like all the jobs in the script are running on > > localhost instead. > > > > > > When I comment out the localhost part from the sies.xml and the > > script, I correctly get them to run on sandyb/westmere queues. > > > > > > In order to verify this, I tried to run a Swift script with three jobs > > each destined to westmere, sandyb and localhost echoing a message and > > printing the hostname. I do find the same behavior in this case too. > > Jobs were all running on localhost and none on the queues. > > > > > > See the above mentioned example bundle with logs and config files > > here: > > > > > > http://www.mcs.anl.gov/~ketan/sites_exp.tgz > > > > > > > > Should run straightaway with ./run.midway and . in PATH. > > > > > > > > > > Feature or bug? > > -- > > Ketan > > > > > > _______________________________________________ > > Swift-devel mailing list > > Swift-devel at ci.uchicago.edu > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From wilde at mcs.anl.gov Sat Apr 20 18:31:21 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Sat, 20 Apr 2013 18:31:21 -0500 (CDT) Subject: [Swift-devel] midway jobs not directed to right site In-Reply-To: Message-ID: <2065170696.4164741.1366500681264.JavaMail.root@mcs.anl.gov> This sounds like a bug, or at the very least a fixable deficiency. - changing just the staging method should not affect where the jobs run - if the coaster provider is smart enough to boot multiple pools of workers, it should similarly be able to associate each pool with a unique service. Maybe the way to ask for that is instead of local:slurm do ssh-cl:slurm. Then it will start multiple separate coaster service JVMs on the loca host using ssh-cl (I *think*). Mike ----- Original Message ----- > From: "Ketan Maheshwari" > To: "Mihael Hategan" > Cc: "Swift Devel" > Sent: Saturday, April 20, 2013 10:48:33 AM > Subject: Re: [Swift-devel] midway jobs not directed to right site > > > > Thanks Mihael! Works after turning off provider staging and setting > filesystem provider in each of the queue pools. > > > > On Sat, Apr 20, 2013 at 12:59 AM, Mihael Hategan < > hategan at mcs.anl.gov > wrote: > > > They probably go to the right site, but all of those sites point to > whatever coaster service runs on localhost of which there is only one > (once one is started, whenever you try to run a coaster job on > localhost, the already started service will be used). > > So you might want to change the last site to use the local provider > instead. > > Mihael > > > > On Fri, 2013-04-19 at 20:32 -0500, Ketan Maheshwari wrote: > > Hi, > > > > I am trying to run a script where some lightweight jobs were > > destined to > > run on localhost while others were to run on sandyb/westmere queues > > in > > midway. > > > > I do this by putting the right handle before the app name in my tc. > > The > > corresponding pool entry being in the sites.xml > > > > However, it looks like all the jobs in the script are running on > > localhost > > instead. > > > > When I comment out the localhost part from the sies.xml and the > > script, I > > correctly get them to run on sandyb/westmere queues. > > > > In order to verify this, I tried to run a Swift script with three > > jobs each > > destined to westmere, sandyb and localhost echoing a message and > > printing > > the hostname. I do find the same behavior in this case too. Jobs > > were all > > running on localhost and none on the queues. > > > > See the above mentioned example bundle with logs and config files > > here: > > > > http://www.mcs.anl.gov/~ketan/sites_exp.tgz > > > > Should run straightaway with ./run.midway and . in PATH. > > > > > > Feature or bug? > > Hi, > > > > > > I am trying to run a script where some lightweight jobs were > > destined > > to run on localhost while others were to run on sandyb/westmere > > queues > > in midway. > > > > > > I do this by putting the right handle before the app name in my tc. > > The corresponding pool entry being in the sites.xml > > > > > > However, it looks like all the jobs in the script are running on > > localhost instead. > > > > > > When I comment out the localhost part from the sies.xml and the > > script, I correctly get them to run on sandyb/westmere queues. > > > > > > In order to verify this, I tried to run a Swift script with three > > jobs > > each destined to westmere, sandyb and localhost echoing a message > > and > > printing the hostname. I do find the same behavior in this case > > too. > > Jobs were all running on localhost and none on the queues. > > > > > > See the above mentioned example bundle with logs and config files > > here: > > > > > > http://www.mcs.anl.gov/~ketan/sites_exp.tgz > > > > > > > > Should run straightaway with ./run.midway and . in PATH. > > > > > > > > > > Feature or bug? > > -- > > Ketan > > > > > > _______________________________________________ > > Swift-devel mailing list > > Swift-devel at ci.uchicago.edu > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > > > > > -- > Ketan > > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > From ketancmaheshwari at gmail.com Sun Apr 21 00:26:44 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Sun, 21 Apr 2013 00:26:44 -0500 Subject: [Swift-devel] turning off Swift's data caching Message-ID: Hi, I am trying to mimic a scenario where an external process is streaming data into a file which is being captured at same or lesser frequency by updating a file periodically which is then cat'ed by Swift into an output file. The Swift script is the catsnsleep with a sleep of 10sec followed by cat of input data. Another process running at a slightly higher frequency does update on the data by echoing random numbers to the data file. >From the experiments so far, it looks like Swift caches the data file and uses it for all app calls. So far, I tried: 1) using external variable type for input data 2) setting caching.algorithm=none in config Neither of the above seem to work. I think, if I could map external type to a disc file, this might work. Any other ideas to turn off input data caching by Swift? Thanks, -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From wozniak at mcs.anl.gov Mon Apr 22 13:40:30 2013 From: wozniak at mcs.anl.gov (Justin M Wozniak) Date: Mon, 22 Apr 2013 13:40:30 -0500 Subject: [Swift-devel] turning off Swift's data caching In-Reply-To: References: Message-ID: <5175841E.8070907@mcs.anl.gov> Normally, Swift variables are write-once. Swift assumes the contents of a file are constant and is allowed to cache. However, you can do this in Swift/T. Swift/T does not stage files, so there is no cache, and there are new constructs for handling external data dependencies. The following script launches compute() every 10 seconds if data.txt has changed. Touch done.txt to exit. import io; import files; import sys; app compute(file f) { "/bin/echo" "compute on:" f; } (int o) check_done(int i) { i => if (file_exists("done.txt")) { o = 1; } else { o = 0; } } global const int interval = 10; main { int i, done; for (i = 0, done = 0; done == 0; i = next_i, done = check_done(i)) { int next_i; i => { if (file_mtime("data.txt") > clock_seconds()-interval) { compute(input_file("data.txt")); } sleep(itof(interval)) => next_i = 1; } } } On 04/21/2013 12:26 AM, Ketan Maheshwari wrote: > Hi, > > I am trying to mimic a scenario where an external process is streaming > data into a file which is being captured at same or lesser frequency > by updating a file periodically which is then cat'ed by Swift into an > output file. -- Justin M Wozniak From tim.g.armstrong at gmail.com Mon Apr 22 14:44:36 2013 From: tim.g.armstrong at gmail.com (Tim Armstrong) Date: Mon, 22 Apr 2013 14:44:36 -0500 Subject: [Swift-devel] turning off Swift's data caching In-Reply-To: <5175841E.8070907@mcs.anl.gov> References: <5175841E.8070907@mcs.anl.gov> Message-ID: Be warned that if you go down the path of having Swift/T file variables based on changing file-system files, then the language doesn't really guarantee any particular behaviour: the behaviour may be surprising. You might be able to get it to do what you want for now: as Justin mentioned, there isn't runtime caching of files at the moment (this of course may change in future). Even in current versions, the optimizer will make the assumption that file contents don't change. I.e. that files are single-assignment. Justin's specific example works in current versions of Swift/T, but I won't guarantee that it will work in future versions as the optimizer gets smarter. There are also subtle changes you can make to the code that will make it cache the results of compute(). This is because, even though it doesn't cache things at runtime, compiler optimizations will assume that the contents of files don't change. In general, the Swift model has to assume that file variables are unchanging, in part because its declarative, but also because we don't really have a strong notion of sequencing or time, so it's not really possible to provide any sort of reasonable semantics to a programmer for mutating files, especially when performance is a concern. I spoke to Justin about this off-list and it sounds like he has a workaround in mind that involves having a sequence number argument to compute() that should be more robust. I did want to send this reply to this list so that others know that "here be dragons" when it comes to dealing with mutating files in Swift/T as well as /K. - Tim On Mon, Apr 22, 2013 at 1:40 PM, Justin M Wozniak wrote: > > Normally, Swift variables are write-once. Swift assumes the contents of > a file are constant and is allowed to cache. > > However, you can do this in Swift/T. Swift/T does not stage files, so > there is no cache, and there are new constructs for handling external > data dependencies. > > The following script launches compute() every 10 seconds if data.txt has > changed. Touch done.txt to exit. > > import io; > import files; > import sys; > > app compute(file f) { > "/bin/echo" "compute on:" f; > } > > (int o) check_done(int i) { > i => > if (file_exists("done.txt")) { o = 1; } > else { o = 0; } > } > > global const int interval = 10; > > main { > int i, done; > for (i = 0, done = 0; done == 0; i = next_i, done = check_done(i)) { > int next_i; > i => { > if (file_mtime("data.txt") > clock_seconds()-interval) { > compute(input_file("data.txt")); > } > sleep(itof(interval)) => next_i = 1; > } > } > } > > > On 04/21/2013 12:26 AM, Ketan Maheshwari wrote: > > Hi, > > > > I am trying to mimic a scenario where an external process is streaming > > data into a file which is being captured at same or lesser frequency > > by updating a file periodically which is then cat'ed by Swift into an > > output file. > > -- > Justin M Wozniak > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > -------------- next part -------------- An HTML attachment was scrubbed... URL: From yadudoc1729 at gmail.com Wed Apr 24 11:45:14 2013 From: yadudoc1729 at gmail.com (Yadu Nand) Date: Wed, 24 Apr 2013 22:15:14 +0530 Subject: [Swift-devel] Moving to Chicago [OT] Message-ID: Hello Everyone, I am moving from India to Chicago this 8th of May. I'm thinking of staying in or near the University for the summer, where I could work from the Computation Institute building, and take the shuttle to Argonne for a few days of the week. As for the first few days, I tried the International House, but they seem to have no rooms available till the 11th. If any of you know a cheap alternative please let me know. Thanks, -Yadu From tim.g.armstrong at gmail.com Wed Apr 24 11:59:21 2013 From: tim.g.armstrong at gmail.com (Tim Armstrong) Date: Wed, 24 Apr 2013 11:59:21 -0500 Subject: [Swift-devel] Moving to Chicago [OT] In-Reply-To: References: Message-ID: Hi Yadu, The Ramada Inn near the university has a discount for University of Chicago visitors. I booked a room recently for a visitor and it was slightly over $100 a night - so probably not that much more than I-House. It was the best deal I could find that wasn't too far away from the university. It's a bit further away than International House: maybe closer to 1/2 an hour walk, but there is a bus you can catch. - Tim On Wed, Apr 24, 2013 at 11:45 AM, Yadu Nand wrote: > Hello Everyone, > > I am moving from India to Chicago this 8th of May. I'm thinking of staying > in or near the University for the summer, where I could work from the > Computation Institute building, and take the shuttle to Argonne for a few > days of the week. > > As for the first few days, I tried the International House, but they seem > to > have no rooms available till the 11th. If any of you know a cheap > alternative > please let me know. > > Thanks, > -Yadu > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > -------------- next part -------------- An HTML attachment was scrubbed... URL: From yadudoc1729 at gmail.com Wed Apr 24 12:19:04 2013 From: yadudoc1729 at gmail.com (Yadu Nand) Date: Wed, 24 Apr 2013 22:49:04 +0530 Subject: [Swift-devel] Moving to Chicago [OT] In-Reply-To: References: Message-ID: Hi Tim, I just checked with Ramada Inn, and got a room booked for $105 + taxes, which is relatively good rate. Thanks for pointing me there :) -Yadu On Wed, Apr 24, 2013 at 10:29 PM, Tim Armstrong wrote: > Hi Yadu, > The Ramada Inn near the university has a discount for University of > Chicago visitors. I booked a room recently for a visitor and it was > slightly over $100 a night - so probably not that much more than I-House. > It was the best deal I could find that wasn't too far away from the > university. It's a bit further away than International House: maybe closer > to 1/2 an hour walk, but there is a bus you can catch. > > - Tim > > > On Wed, Apr 24, 2013 at 11:45 AM, Yadu Nand wrote: >> >> Hello Everyone, >> >> I am moving from India to Chicago this 8th of May. I'm thinking of staying >> in or near the University for the summer, where I could work from the >> Computation Institute building, and take the shuttle to Argonne for a few >> days of the week. >> >> As for the first few days, I tried the International House, but they seem >> to >> have no rooms available till the 11th. If any of you know a cheap >> alternative >> please let me know. >> >> Thanks, >> -Yadu >> _______________________________________________ >> Swift-devel mailing list >> Swift-devel at ci.uchicago.edu >> https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > -- Yadu Nand B From davidk at ci.uchicago.edu Wed Apr 24 15:51:52 2013 From: davidk at ci.uchicago.edu (David Kelly) Date: Wed, 24 Apr 2013 15:51:52 -0500 (CDT) Subject: [Swift-devel] auto-coaster bootstrap for stampede cluster In-Reply-To: <1617183787.3432082.1366231891105.JavaMail.root@mcs.anl.gov> Message-ID: <1658732124.4895607.1366836712665.JavaMail.root@ci.uchicago.edu> Ketan, I have gram working to Stampede now. Given the restrictions about running swift on the head nodes, I think this is the way to go. I'll add this info to the site guide, but for now here is a quick overview of what's needed. Get a proxy: myproxy-logon -l username -s myproxy.teragrid.org Make sure you have GLOBUS_HOSTNAME and GLOBUS_TCP_PORT_RANGE defined correctly. Use something like this for your sites .xml (with work directory, project, and throttle adjusted as needed) --- 16 16 3600 00:05:00 100 100 normal 1 1 TG-EAR130015 .3199 10000 /scratch/01503/davidkel --- You'll also need the latest version of Swift from SVN. Swift was setting some invalid gram RSL attributes that were causing jobs to fail. I added a check to verify only valid attributes get set now. I've tested this with a simple swift script that calls /bin/hostname and it ran across multiple Stampede nodes. I haven't tested it with any larger applications yet - please let me know if you run into any problems with it. Thanks, David ----- Original Message ----- From: "Michael Wilde" To: "Swift Devel" Sent: Wednesday, April 17, 2013 3:51:31 PM Subject: [Swift-devel] auto-coaster bootstrap for stampede cluster I'm moving this topic to swift-devel, so others, in particular Mihael, can weigh in. - Mike ----- Forwarded Message ----- From: "David Kelly" To: "Ketan Maheshwari" Cc: "Wilde" Sent: Wednesday, April 17, 2013 3:45:30 PM Subject: Fwd: auto-coaster bootstrap for stamped Hey Ketan, Mike mentioned that you were interested in running remotely to Stampede via ssh-cl. Normally we could use ssh-cl like any other site, but the problem we ran into here is that we can't run Swift on the stampede head node. We need to ssh-cl AND also start swift on a remote worker node, which is a setup that hasn't been tested very much. I believe you've used start-coaster-service before when we were running on ec2. You can this configuration for Stampede too. Modify coaster-service.conf to set WORKER_NODE=slurm, WORKER_RELAY_HOST=stampede.tacc.utexas.edu, and it will generate a slurm script, scp it to stampede, and remotely start swift on a worker node. I'll see if I can find an example config file for this. With automatic coaters it's a bit more complicated and completely untested as far as I know. You may be able to use gram2. This worked on Ranger, but haven't tried yet on Stampede. Mike mentioned in the email below you may be able to change the ssh-cl provider to add some kind of prefix command (srun). Maybe you can modify your PATH so the 'ssh' command is actually a wrapper you created and does something sneaky. You may also be able to add a prefix command to cog/modules/provider-coaster/resources/bootstrap.sh. Hopefully this can help you get started - let me know if any of this works for you, curious to see how we can get it working well. David ----- Forwarded Message ----- From: "Michael Wilde" To: "David Kelly" Sent: Tuesday, April 16, 2013 10:59:22 AM Subject: auto-coaster bootstrap for stamped was: Re: Another item for the to-do list David, thanks for the details. Im wondering, for systems like stampede, could automatic coasters work to it (eg from swift.rcc) by adding a sinteractive or srun command into the middle of the ssh command generated by the ssh-cl parameter? ie instead of doing ssh -sshargsgere auto-boostrap-coaster-stuff-here.sh do: ssh -sshargsgere srun auto-boostrap-coaster-stuff-here.sh ? > This is the only mode that I've been able to test on Stampede so far. > I will experiment more the others when Stampede is back up. Others meaning GRAM? Perhaps using myproxy-logon? That *should* work out of the box but we've not tested GRAM in ages so it probably doesnt. Lets keep this lower on the prio list. I just want to be sure we have a ticket for this. Please create one if not - thanks. _______________________________________________ Swift-devel mailing list Swift-devel at ci.uchicago.edu https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel -------------- next part -------------- An HTML attachment was scrubbed... URL: From ketancmaheshwari at gmail.com Wed Apr 24 17:10:34 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Wed, 24 Apr 2013 17:10:34 -0500 Subject: [Swift-devel] auto-coaster bootstrap for stampede cluster In-Reply-To: <1658732124.4895607.1366836712665.JavaMail.root@ci.uchicago.edu> References: <1617183787.3432082.1366231891105.JavaMail.root@mcs.anl.gov> <1658732124.4895607.1366836712665.JavaMail.root@ci.uchicago.edu> Message-ID: Thanks David. This sounds very useful for Stampede. I will try it for the remaining VASP runs on Stampede. On Wed, Apr 24, 2013 at 3:51 PM, David Kelly wrote: > Ketan, > > I have gram working to Stampede now. Given the restrictions about running > swift on the head nodes, I think this is the way to go. I'll add this > info to the site guide, but for now here is a quick overview of what's > needed. > > Get a proxy: myproxy-logon -l username -s myproxy.teragrid.org > > Make sure you have GLOBUS_HOSTNAME and GLOBUS_TCP_PORT_RANGE defined > correctly. > > Use something like this for your sites.xml (with work directory, project, > and throttle adjusted as needed) > --- > > > > > 16 > 16 > 3600 > 00:05:00 > 100 > 100 > normal > 1 > 1 > TG-EAR130015 > .3199 > 10000 > /scratch/01503/davidkel > > > --- > > You'll also need the latest version of Swift from SVN. Swift was setting > some invalid gram RSL attributes that were causing jobs to fail. I added a > check to verify only valid attributes get set now. I've tested this with a > simple swift script that calls /bin/hostname and it ran across multiple > Stampede nodes. I haven't tested it with any larger applications yet - > please let me know if you run into any problems with it. > > Thanks, > David > > ------------------------------ > > *From: *"Michael Wilde" > *To: *"Swift Devel" > *Sent: *Wednesday, April 17, 2013 3:51:31 PM > *Subject: *[Swift-devel] auto-coaster bootstrap for stampede cluster > > > > I'm moving this topic to swift-devel, so others, in particular Mihael, can > weigh in. > > - Mike > > ----- Forwarded Message ----- > From: "David Kelly" > To: "Ketan Maheshwari" > Cc: "Wilde" > Sent: Wednesday, April 17, 2013 3:45:30 PM > Subject: Fwd: auto-coaster bootstrap for stamped > > Hey Ketan, > > Mike mentioned that you were interested in running remotely to Stampede > via ssh-cl. Normally we could use ssh-cl like any other site, but the > problem we ran into here is that we can't run Swift on the stampede head > node. We need to ssh-cl AND also start swift on a remote worker node, which > is a setup that hasn't been tested very much. > > I believe you've used start-coaster-service before when we were running on > ec2. You can this configuration for Stampede too. Modify > coaster-service.conf to set WORKER_NODE=slurm, WORKER_RELAY_HOST= > stampede.tacc.utexas.edu, and it will generate a slurm script, scp it to > stampede, and remotely start swift on a worker node. I'll see if I can find > an example config file for this. > > With automatic coaters it's a bit more complicated and completely untested > as far as I know. > > You may be able to use gram2. This worked on Ranger, but haven't tried yet > on Stampede. > Mike mentioned in the email below you may be able to change the ssh-cl > provider to add some kind of prefix command (srun). > Maybe you can modify your PATH so the 'ssh' command is actually a wrapper > you created and does something sneaky. > You may also be able to add a prefix command to > cog/modules/provider-coaster/resources/bootstrap.sh. > > Hopefully this can help you get started - let me know if any of this works > for you, curious to see how we can get it working well. > > David > > ----- Forwarded Message ----- > > > From: "Michael Wilde" > To: "David Kelly" > Sent: Tuesday, April 16, 2013 10:59:22 AM > Subject: auto-coaster bootstrap for stamped > > > was: Re: Another item for the to-do list > > David, thanks for the details. > > Im wondering, for systems like stampede, could automatic coasters work to > it (eg from swift.rcc) by adding a sinteractive or srun command into the > middle of the ssh command generated by the ssh-cl parameter? > > ie instead of doing ssh -sshargsgere auto-boostrap-coaster-stuff-here.sh > do: ssh -sshargsgere srun auto-boostrap-coaster-stuff-here.sh > > ? > > > This is the only mode that I've been able to test on Stampede so far. > > I will experiment more the others when Stampede is back up. > > Others meaning GRAM? Perhaps using myproxy-logon? That *should* work out > of the box but we've not tested GRAM in ages so it probably doesnt. > > Lets keep this lower on the prio list. I just want to be sure we have a > ticket for this. Please create one if not - thanks. > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: From wilde at mcs.anl.gov Thu Apr 25 09:25:04 2013 From: wilde at mcs.anl.gov (Michael Wilde) Date: Thu, 25 Apr 2013 09:25:04 -0500 Subject: [Swift-devel] auto-coaster bootstrap for stampede cluster In-Reply-To: <1658732124.4895607.1366836712665.JavaMail.root@ci.uchicago.edu> References: <1617183787.3432082.1366231891105.JavaMail.root@mcs.anl.gov> <1658732124.4895607.1366836712665.JavaMail.root@ci.uchicago.edu> Message-ID: David, this sounds great - nice work. Can you test with multiple, mixed sites and provider and gridftp staging? Try e.g. Stampede+trestles(+midway+beagle+kraken) Also gt2:slurm:slurm might work well. Please add this all to the site guide (ideally with a diagram). Mihael, how hard would it be to make ssh-cl:slurm:slurm to work? I.e. start the coaster service ond the remote site as a slurm job instaed of on the login host, which is the objective of this configuration. Very cool. - Mike On 4/24/13, David Kelly wrote: > Ketan, > > > I have gram working to Stampede now. Given the restrictions about running > swift on the head nodes, I think this is the way to go. I'll add this info > to the site guide, but for now here is a quick overview of what's needed. > > > Get a proxy: myproxy-logon -l username -s myproxy.teragrid.org > > > Make sure you have GLOBUS_HOSTNAME and GLOBUS_TCP_PORT_RANGE defined > correctly. > > > Use something like this for your sites .xml (with work directory, project, > and throttle adjusted as needed) > --- > > > > > url="login5.stampede.tacc.utexas.edu:2119/jobmanager-slurm"/> > url="gsiftp://gridftp.stampede.tacc.utexas.edu:2811"/> > 16 > 16 > 3600 > 00:05:00 > 100 > 100 > normal > 1 > 1 > TG-EAR130015 > .3199 > 10000 > /scratch/01503/davidkel > > > --- > > > You'll also need the latest version of Swift from SVN. Swift was setting > some invalid gram RSL attributes that were causing jobs to fail. I added a > check to verify only valid attributes get set now. I've tested this with a > simple swift script that calls /bin/hostname and it ran across multiple > Stampede nodes. I haven't tested it with any larger applications yet - > please let me know if you run into any problems with it. > > > Thanks, > David > ----- Original Message ----- > > > From: "Michael Wilde" > To: "Swift Devel" > Sent: Wednesday, April 17, 2013 3:51:31 PM > Subject: [Swift-devel] auto-coaster bootstrap for stampede cluster > > > I'm moving this topic to swift-devel, so others, in particular Mihael, can > weigh in. > > - Mike > > ----- Forwarded Message ----- > From: "David Kelly" > To: "Ketan Maheshwari" > Cc: "Wilde" > Sent: Wednesday, April 17, 2013 3:45:30 PM > Subject: Fwd: auto-coaster bootstrap for stamped > > Hey Ketan, > > Mike mentioned that you were interested in running remotely to Stampede via > ssh-cl. Normally we could use ssh-cl like any other site, but the problem we > ran into here is that we can't run Swift on the stampede head node. We need > to ssh-cl AND also start swift on a remote worker node, which is a setup > that hasn't been tested very much. > > I believe you've used start-coaster-service before when we were running on > ec2. You can this configuration for Stampede too. Modify > coaster-service.conf to set WORKER_NODE=slurm, > WORKER_RELAY_HOST=stampede.tacc.utexas.edu, and it will generate a slurm > script, scp it to stampede, and remotely start swift on a worker node. I'll > see if I can find an example config file for this. > > With automatic coaters it's a bit more complicated and completely untested > as far as I know. > > You may be able to use gram2. This worked on Ranger, but haven't tried yet > on Stampede. > Mike mentioned in the email below you may be able to change the ssh-cl > provider to add some kind of prefix command (srun). > Maybe you can modify your PATH so the 'ssh' command is actually a wrapper > you created and does something sneaky. > You may also be able to add a prefix command to > cog/modules/provider-coaster/resources/bootstrap.sh. > > Hopefully this can help you get started - let me know if any of this works > for you, curious to see how we can get it working well. > > David > > ----- Forwarded Message ----- > > > From: "Michael Wilde" > To: "David Kelly" > Sent: Tuesday, April 16, 2013 10:59:22 AM > Subject: auto-coaster bootstrap for stamped > > > was: Re: Another item for the to-do list > > David, thanks for the details. > > Im wondering, for systems like stampede, could automatic coasters work to it > (eg from swift.rcc) by adding a sinteractive or srun command into the middle > of the ssh command generated by the ssh-cl parameter? > > ie instead of doing ssh -sshargsgere auto-boostrap-coaster-stuff-here.sh > do: ssh -sshargsgere srun auto-boostrap-coaster-stuff-here.sh > > ? > >> This is the only mode that I've been able to test on Stampede so far. >> I will experiment more the others when Stampede is back up. > > Others meaning GRAM? Perhaps using myproxy-logon? That *should* work out of > the box but we've not tested GRAM in ages so it probably doesnt. > > Lets keep this lower on the prio list. I just want to be sure we have a > ticket for this. Please create one if not - thanks. > > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > > -- Sent from my mobile device From ketancmaheshwari at gmail.com Sun Apr 28 00:04:10 2013 From: ketancmaheshwari at gmail.com (Ketan Maheshwari) Date: Sun, 28 Apr 2013 00:04:10 -0500 Subject: [Swift-devel] auto-coaster bootstrap for stampede cluster In-Reply-To: References: <1617183787.3432082.1366231891105.JavaMail.root@mcs.anl.gov> <1658732124.4895607.1366836712665.JavaMail.root@ci.uchicago.edu> Message-ID: I tested this successfully from communicado into Stampede. One tricky issue I fell into was that ibrun on Stampede checks for keys in ~/.ssh originally generated by stampede at the time of first login. I replaced them with my own keypair that I use on other machines. This was causing the jobs to subtly fail without any explicit error message on stderr of Swift nor on the gram log. The issue was resolved after digging into Stampede manual and confirming with a similar buried post on xsede forum. Thanks, Ketan On Thu, Apr 25, 2013 at 9:25 AM, Michael Wilde wrote: > David, this sounds great - nice work. > > Can you test with multiple, mixed sites and provider and gridftp > staging? Try e.g. Stampede+trestles(+midway+beagle+kraken) > > Also gt2:slurm:slurm might work well. > > Please add this all to the site guide (ideally with a diagram). > > Mihael, how hard would it be to make ssh-cl:slurm:slurm to work? I.e. > start the coaster service ond the remote site as a slurm job instaed > of on the login host, which is the objective of this configuration. > > Very cool. > > - Mike > > On 4/24/13, David Kelly wrote: > > Ketan, > > > > > > I have gram working to Stampede now. Given the restrictions about running > > swift on the head nodes, I think this is the way to go. I'll add this > info > > to the site guide, but for now here is a quick overview of what's needed. > > > > > > Get a proxy: myproxy-logon -l username -s myproxy.teragrid.org > > > > > > Make sure you have GLOBUS_HOSTNAME and GLOBUS_TCP_PORT_RANGE defined > > correctly. > > > > > > Use something like this for your sites .xml (with work directory, > project, > > and throttle adjusted as needed) > > --- > > > > > > > > > > > url="login5.stampede.tacc.utexas.edu:2119/jobmanager-slurm"/> > > > url="gsiftp://gridftp.stampede.tacc.utexas.edu:2811"/> > > 16 > > 16 > > 3600 > > 00:05:00 > > 100 > > 100 > > normal > > 1 > > 1 > > TG-EAR130015 > > .3199 > > 10000 > > /scratch/01503/davidkel > > > > > > --- > > > > > > You'll also need the latest version of Swift from SVN. Swift was setting > > some invalid gram RSL attributes that were causing jobs to fail. I added > a > > check to verify only valid attributes get set now. I've tested this with > a > > simple swift script that calls /bin/hostname and it ran across multiple > > Stampede nodes. I haven't tested it with any larger applications yet - > > please let me know if you run into any problems with it. > > > > > > Thanks, > > David > > ----- Original Message ----- > > > > > > From: "Michael Wilde" > > To: "Swift Devel" > > Sent: Wednesday, April 17, 2013 3:51:31 PM > > Subject: [Swift-devel] auto-coaster bootstrap for stampede cluster > > > > > > I'm moving this topic to swift-devel, so others, in particular Mihael, > can > > weigh in. > > > > - Mike > > > > ----- Forwarded Message ----- > > From: "David Kelly" > > To: "Ketan Maheshwari" > > Cc: "Wilde" > > Sent: Wednesday, April 17, 2013 3:45:30 PM > > Subject: Fwd: auto-coaster bootstrap for stamped > > > > Hey Ketan, > > > > Mike mentioned that you were interested in running remotely to Stampede > via > > ssh-cl. Normally we could use ssh-cl like any other site, but the > problem we > > ran into here is that we can't run Swift on the stampede head node. We > need > > to ssh-cl AND also start swift on a remote worker node, which is a setup > > that hasn't been tested very much. > > > > I believe you've used start-coaster-service before when we were running > on > > ec2. You can this configuration for Stampede too. Modify > > coaster-service.conf to set WORKER_NODE=slurm, > > WORKER_RELAY_HOST=stampede.tacc.utexas.edu, and it will generate a slurm > > script, scp it to stampede, and remotely start swift on a worker node. > I'll > > see if I can find an example config file for this. > > > > With automatic coaters it's a bit more complicated and completely > untested > > as far as I know. > > > > You may be able to use gram2. This worked on Ranger, but haven't tried > yet > > on Stampede. > > Mike mentioned in the email below you may be able to change the ssh-cl > > provider to add some kind of prefix command (srun). > > Maybe you can modify your PATH so the 'ssh' command is actually a wrapper > > you created and does something sneaky. > > You may also be able to add a prefix command to > > cog/modules/provider-coaster/resources/bootstrap.sh. > > > > Hopefully this can help you get started - let me know if any of this > works > > for you, curious to see how we can get it working well. > > > > David > > > > ----- Forwarded Message ----- > > > > > > From: "Michael Wilde" > > To: "David Kelly" > > Sent: Tuesday, April 16, 2013 10:59:22 AM > > Subject: auto-coaster bootstrap for stamped > > > > > > was: Re: Another item for the to-do list > > > > David, thanks for the details. > > > > Im wondering, for systems like stampede, could automatic coasters work > to it > > (eg from swift.rcc) by adding a sinteractive or srun command into the > middle > > of the ssh command generated by the ssh-cl parameter? > > > > ie instead of doing ssh -sshargsgere auto-boostrap-coaster-stuff-here.sh > > do: ssh -sshargsgere srun auto-boostrap-coaster-stuff-here.sh > > > > ? > > > >> This is the only mode that I've been able to test on Stampede so far. > >> I will experiment more the others when Stampede is back up. > > > > Others meaning GRAM? Perhaps using myproxy-logon? That *should* work out > of > > the box but we've not tested GRAM in ages so it probably doesnt. > > > > Lets keep this lower on the prio list. I just want to be sure we have a > > ticket for this. Please create one if not - thanks. > > > > _______________________________________________ > > Swift-devel mailing list > > Swift-devel at ci.uchicago.edu > > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > > > > > > > > > > -- > Sent from my mobile device > _______________________________________________ > Swift-devel mailing list > Swift-devel at ci.uchicago.edu > https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel > -- Ketan -------------- next part -------------- An HTML attachment was scrubbed... URL: