[Swift-devel] PBS time-exceeds-queue conflict yields cryptic error

[Michael Wilde]

A low prio issue:

When I ask for more time than the selected PBS queue allows, I get a 
cryptic error. The fact that this condition yields a PBS error is known 
and has been discussed on the list.

Is it tracked as bug 133, or does that refer to exceed allotted time at 
runtime? If so, I can add this note; else I can file a new bug.

In my case, I gave:

<config>
<pool handle="teraport" >
   <profile namespace="globus" key="queue">fast</profile>
   <profile namespace="globus" key="maxwalltime">05:00:00</profile>
   <execution provider="pbs" url="none" />
   <gridftp  url="local://localhost" />
   <workdirectory>/home/wilde/swiftwork</workdirectory>
</pool>
</config>

my error was asking for 5 hours (05:00:00) instead of 5 minutes.

I got the un-helpful error:

tp$ swift -tc.file tc.data -sites.file teraport.xml floop.swift
Swift svn swift-r2631 cog-r2306

RunID: 20090306-1822-cerzn3y8
Progress:
Failed to transfer wrapper log from floop-20090306-1822-cerzn3y8/info/0 
on teraport
Execution failed:
	Failed to transfer wrapper log from floop-20090306-1822-cerzn3y8/info/r 
on teraport
Failed to transfer wrapper log from floop-20090306-1822-cerzn3y8/info/y 
on teraport
Progress:  Submitting:7 Failed:3
Exception in echo:
Arguments: [42]
Host: teraport
Directory: floop-20090306-1822-cerzn3y8/jobs/0/echo-0hikdk7j
stderr.txt:

stdout.txt:

----

Caused by:
	Cannot submit job: Could not submit job (qsub reported an exit code of 
188). no error output
Failed to transfer wrapper log from floop-20090306-1822-cerzn3y8/info/s 
on teraport
Failed to transfer wrapper log from floop-20090306-1822-cerzn3y8/info/q 
on teraport
tp$ fg