<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Tahoma
}
--></style>
</head>
<body class='hmmessage'><div dir='ltr'>
I already did that in the revision I have, and the tests passed every iteration; I am going to update my working copies to the latest svn code, and I'll run it again.<div>By the way, I put sites.template.xml back, and ran the command jstack</div><div><br></div><div>This is the output:</div><div><br></div><div><div>$ jstack -l 2791</div><div>2791: Unable to open socket file: target process not responding or HotSpot VM not loaded</div><div>The -F option can be used when the target process is not responding</div><div>$ jstack -F 2791</div><div>Attaching to process ID 2791, please wait...</div><div>sun.jvm.hotspot.debugger.NoSuchSymbolException: Could not find symbol "gHotSpotVMTypes" in any of the known library names (libjvm.so, libjvm_g.so, gamma_g)</div><div> at sun.jvm.hotspot.HotSpotTypeDataBase.lookupInProcess(HotSpotTypeDataBase.java:389)</div><div> at sun.jvm.hotspot.HotSpotTypeDataBase.readVMTypes(HotSpotTypeDataBase.java:104)</div><div> at sun.jvm.hotspot.HotSpotTypeDataBase.<init>(HotSpotTypeDataBase.java:85)</div><div> at sun.jvm.hotspot.bugspot.BugSpotAgent.setupVM(BugSpotAgent.java:568)</div><div> at sun.jvm.hotspot.bugspot.BugSpotAgent.go(BugSpotAgent.java:494)</div><div> at sun.jvm.hotspot.bugspot.BugSpotAgent.attach(BugSpotAgent.java:332)</div><div> at sun.jvm.hotspot.tools.Tool.start(Tool.java:163)</div><div> at sun.jvm.hotspot.tools.JStack.main(JStack.java:86)</div><div> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)</div><div> at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)</div><div> at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)</div><div> at java.lang.reflect.Method.invoke(Method.java:597)</div><div> at sun.tools.jstack.JStack.runJStackTool(JStack.java:118)</div><div> at sun.tools.jstack.JStack.main(JStack.java:84)</div><div>Debugger attached successfully.</div><div>jstack requires a java VM process/core!</div></div><div><br></div><div><br></div><div><div>> Subject: RE: [Swift-devel] Swift unresponsive while using local provider.<br>> From: hategan@mcs.anl.gov<br>> To: alberto_chavez@live.com<br>> CC: davidkelly999@gmail.com; swift-devel@ci.uchicago.edu<br>> Date: Fri, 17 Jun 2011 17:19:46 -0700<br>> <br>> Right. Now run it 100 more times (make a loop in a shell script) and see<br>> if none of those deadlock.<br>> <br>> Then update to the latest svn code, re-compile and run the script 100<br>> more times. See if it deadlocks then.<br>> <br>> On Fri, 2011-06-17 at 19:13 -0500, Alberto Chavez wrote:<br>> > No, is the same one.<br>> > But I had a sites.template.xml file in that directory, which contained<br>> > that information; as soon as I removed sites.template.xml from my<br>> > directory, the script worked just fine.<br>> > <br>> > <br>> > type messagefile;<br>> > app (messagefile t) greeting (string s[]) {<br>> > echo s[0] s[1] s[2] stdout=@filename(t);<br>> > }<br>> > messagefile outfile <"q5out.txt">;<br>> > string words[] = ["how","are","you"];<br>> > outfile = greeting(words);<br>> > <br>> > <br>> > <br>> > <br>> > <br>> > > Subject: RE: [Swift-devel] Swift unresponsive while using local<br>> > provider.<br>> > > From: hategan@mcs.anl.gov<br>> > > To: alberto_chavez@live.com<br>> > > CC: davidkelly999@gmail.com; swift-devel@ci.uchicago.edu<br>> > > Date: Fri, 17 Jun 2011 17:09:52 -0700<br>> > > <br>> > > I'm sorry, but I don't follow. Is there a new error?<br>> > > <br>> > > <br>> > > On Fri, 2011-06-17 at 18:55 -0500, Alberto Chavez wrote:<br>> > > > sites.template.xml is producing this error, as soon as I remove<br>> > the<br>> > > > file from the directory, the error goes away as well.<br>> > > > These are the contents of such file:<br>> > > > <br>> > > > <config><br>> > > > <pool handle="localhost"><br>> > > > <filesystem provider="local" /><br>> > > > <execution provider="coaster" jobmanager="local:local"/><br>> > > > <profile namespace="globus"<br>> > > > key="internalHostname">127.0.0.1</profile><br>> > > > <profile namespace="karajan" key="jobthrottle">1000</profile><br>> > > > <profile namespace="karajan" key="initialScore">10000</profile><br>> > > > <profile namespace="globus" key="jobsPerNode">4</profile><br>> > > > <profile namespace="globus" key="slots">8</profile><br>> > > > <profile namespace="globus" key="maxTime">1000</profile><br>> > > > <profile namespace="globus" key="nodeGranularity">1</profile><br>> > > > <profile namespace="globus" key="maxNodes">4</profile><br>> > > > <workdirectory>/tmp</workdirectory><br>> > > > </pool><br>> > > > </config><br>> > > > <br>> > > > <br>> > > > -Alberto<br>> > > > <br>> > > > <br>> > > > > Subject: Re: [Swift-devel] Swift unresponsive while using local<br>> > > > provider.<br>> > > > > From: hategan@mcs.anl.gov<br>> > > > > To: davidkelly999@gmail.com<br>> > > > > Date: Fri, 17 Jun 2011 12:56:25 -0700<br>> > > > > CC: swift-devel@ci.uchicago.edu<br>> > > > > <br>> > > > > do "jstack -l <pid_of_swift_java_process>" whenever it happens<br>> > and<br>> > > > send<br>> > > > > the output.<br>> > > > > <br>> > > > > <br>> > > > > <br>> > > > > On Fri, 2011-06-17 at 14:48 -0500, David Kelly wrote:<br>> > > > > > I saw similar things on my laptop (4 gb ram) this weekend when<br>> > I<br>> > > > was<br>> > > > > > testing the galaxy demo scripts using the local provider. I<br>> > was<br>> > > > using<br>> > > > > > trunk. In the output I would see things like "no activity for<br>> > 10s"<br>> > > > and<br>> > > > > > it just would sit there and do nothing until I manually killed<br>> > it.<br>> > > > But<br>> > > > > > most of the time it would work fine. I wrote a little shell<br>> > script<br>> > > > > > that would repeatedly run it until it hung. Then I was talking<br>> > to<br>> > > > Jon<br>> > > > > > about this and he saw something similar with his montage work.<br>> > He<br>> > > > > > thought it might be related to a configuration issue - that<br>> > either<br>> > > > > > wrapper.parameter.mode=files or status.mode=provider should be<br>> > > > set.<br>> > > > > > <br>> > > > > > I can send my scripts as well if you need some help in<br>> > tracking<br>> > > > this<br>> > > > > > down.<br>> > > > > > <br>> > > > > > David<br>> > > > > > <br>> > > > > > On Fri, Jun 17, 2011 at 2:38 PM, Michael Wilde<br>> > <wilde@mcs.anl.gov><br>> > > > > > wrote:<br>> > > > > > Alberto, how long are you letting it run for, and under what<br>> > > > > > environment? if you are running on your laptop, how much RAM<br>> > > > > > do you have? Its possible that you are seeing paging delays<br>> > > > > > if you are running the Swift Java app with too little memory.<br>> > > > > > <br>> > > > > > <br>> > > > > > Also, are you running trunk or 0.92.1? You should compare the<br>> > > > > > two.<br>> > > > > > <br>> > > > > > <br>> > > > > > Its *possible* that this simple test is hanging under recent<br>> > > > > > trunk mods, but its more likely that this is some kind of<br>> > > > > > resource shortage.<br>> > > > > > <br>> > > > > > <br>> > > > > > Can you run this on one of the Swift lab machines bridled or<br>> > > > > > communcado, or better yet on the MCS compute servers, or a<br>> > > > > > PADS worker node (which you can get with qsub -I on pads)?<br>> > > > > > <br>> > > > > > <br>> > > > > > Look at Swift under the "top" command to see if Swift is<br>> > > > > > running and slow, or is hung.<br>> > > > > > <br>> > > > > > <br>> > > > > > Stop by and we can discuss in more detail.<br>> > > > > > <br>> > > > > > <br>> > > > > > - Mike<br>> > > > > > <br>> > > > > > <br>> > > > > > <br>> > > > > > ______________________________________________________________<br>> > > > > > <br>> > > > > > When I run the following SwiftScript using suite.sh,<br>> > > > > > the report shows an odd behavior, most of the time it<br>> > > > > > times out, but once in a while it passes, however this<br>> > > > > > outcome is completely random, since sometimes that<br>> > > > > > test has passed 3 times in a row, and all of the<br>> > > > > > sudden it fails.<br>> > > > > > This is my script:<br>> > > > > > <br>> > > > > > <br>> > > > > > type messagefile;<br>> > > > > > <br>> > > > > > <br>> > > > > > app (messagefile t) greeting (string s[]) {<br>> > > > > > echo s[0] s[1] s[2] stdout=@filename(t);<br>> > > > > > }<br>> > > > > > <br>> > > > > > <br>> > > > > > messagefile outfile <"q5out.txt">;<br>> > > > > > <br>> > > > > > <br>> > > > > > string words[] = ["how","are","you"];<br>> > > > > > <br>> > > > > > <br>> > > > > > outfile = greeting(words);<br>> > > > > > <br>> > > > > > <br>> > > > > > <br>> > > > > > <br>> > > > > > <br>> > > > > > <br>> > > > > > <br>> > > > > > Swift.properties contents:<br>> > > > > > <br>> > > > > > <br>> > > > > > $ cat swift.properties <br>> > > > > > wrapperlog.always.transfer=true<br>> > > > > > sitedir.keep=true<br>> > > > > > execution.retries=0<br>> > > > > > lazy.errors=false<br>> > > > > > status.mode=provider<br>> > > > > > use.provider.staging=false<br>> > > > > > provider.staging.pin.swiftfiles=false<br>> > > > > > <br>> > > > > > <br>> > > > > > Sites.template.xml contents:<br>> > > > > > <br>> > > > > > <br>> > > > > > $ cat sites.template.xml <br>> > > > > > <config><br>> > > > > > <pool handle="localhost"><br>> > > > > > <filesystem provider="local" /><br>> > > > > > <execution provider="coaster"<br>> > > > > > jobmanager="local:local"/><br>> > > > > > <profile namespace="globus"<br>> > > > > > key="internalHostname">127.0.0.1</profile><br>> > > > > > <profile namespace="karajan"<br>> > > > > > key="jobthrottle">1000</profile><br>> > > > > > <profile namespace="karajan"<br>> > > > > > key="initialScore">10000</profile><br>> > > > > > <profile namespace="globus"<br>> > > > > > key="jobsPerNode">4</profile><br>> > > > > > <profile namespace="globus"<br>> > > > > > key="slots">8</profile><br>> > > > > > <profile namespace="globus"<br>> > > > > > key="maxTime">1000</profile><br>> > > > > > <profile namespace="globus"<br>> > > > > > key="nodeGranularity">1</profile><br>> > > > > > <profile namespace="globus"<br>> > > > > > key="maxNodes">4</profile><br>> > > > > > <workdirectory>/tmp</workdirectory><br>> > > > > > </pool><br>> > > > > > </config><br>> > > > > > <br>> > > > > > <br>> > > > > > -Alberto<br>> > > > > > <br>> > > > > > <br>> > > > > > _______________________________________________<br>> > > > > > Swift-devel mailing list<br>> > > > > > Swift-devel@ci.uchicago.edu<br>> > > > > > http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel<br>> > > > > > <br>> > > > > > <br>> > > > > > <br>> > > > > > -- <br>> > > > > > Michael Wilde<br>> > > > > > Computation Institute, University of Chicago<br>> > > > > > Mathematics and Computer Science Division<br>> > > > > > Argonne National Laboratory<br>> > > > > > <br>> > > > > > <br>> > > > > > <br>> > > > > > _______________________________________________<br>> > > > > > Swift-devel mailing list<br>> > > > > > Swift-devel@ci.uchicago.edu<br>> > > > > > http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel<br>> > > > > > <br>> > > > > > <br>> > > > > > _______________________________________________<br>> > > > > > Swift-devel mailing list<br>> > > > > > Swift-devel@ci.uchicago.edu<br>> > > > > > http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel<br>> > > > > <br>> > > > > <br>> > > > > _______________________________________________<br>> > > > > Swift-devel mailing list<br>> > > > > Swift-devel@ci.uchicago.edu<br>> > > > > http://mail.ci.uchicago.edu/mailman/listinfo/swift-devel<br>> > > > <br>> > > <br>> > > <br>> > <br>> <br>> <br></div></div> </div></body>
</html>