February 2013 - swift-bugs - mailman3.cels.anl.gov

[Bug 96] document writing mappers
by bugzilla-daemon＠mcs.anl.gov 21 Feb '13

21 Feb '13

https://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=96 ketan <ketan(a)mcs.anl.gov> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |FIXED --- Comment #1 from ketan <ketan(a)mcs.anl.gov> 2013-02-21 15:18:01 --- Marking it done as I see the mappers documentation reasonably rounded in userguide. -- Configure bugmail: https://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

1 0

[Bug 74] Campaign: Technology transfer to the user
by bugzilla-daemon＠mcs.anl.gov 16 Feb '13

16 Feb '13

https://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=74 Mihael Hategan <hategan(a)mcs.anl.gov> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |hategan(a)mcs.anl.gov Resolution| |EXPIRED -- Configure bugmail: https://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

1 0

[Bug 74] Campaign: Technology transfer to the user
by bugzilla-daemon＠mcs.anl.gov 16 Feb '13

16 Feb '13

https://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=74 Bug 74 depends on bug 73, which changed state. Bug 73 Summary: Campaign: performance improvements for MolDyn workflow http://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=73 What |Old Value |New Value ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |EXPIRED -- Configure bugmail: https://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

1 0

[Bug 73] Campaign: performance improvements for MolDyn workflow
by bugzilla-daemon＠mcs.anl.gov 16 Feb '13

16 Feb '13

https://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=73 Mihael Hategan <hategan(a)mcs.anl.gov> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |hategan(a)mcs.anl.gov Resolution| |EXPIRED -- Configure bugmail: https://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

1 0

[Bug 73] Campaign: performance improvements for MolDyn workflow
by bugzilla-daemon＠mcs.anl.gov 16 Feb '13

16 Feb '13

https://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=73 Bug 73 depends on bug 72, which changed state. Bug 72 Summary: Campaign for scaling wf up to 244 molecules http://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=72 What |Old Value |New Value ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |EXPIRED -- Configure bugmail: https://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

1 0

[Bug 74] Campaign: Technology transfer to the user
by bugzilla-daemon＠mcs.anl.gov 16 Feb '13

16 Feb '13

https://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=74 Bug 74 depends on bug 72, which changed state. Bug 72 Summary: Campaign for scaling wf up to 244 molecules http://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=72 What |Old Value |New Value ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution| |EXPIRED -- Configure bugmail: https://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

1 0

[Bug 72] Campaign for scaling wf up to 244 molecules
by bugzilla-daemon＠mcs.anl.gov 16 Feb '13

16 Feb '13

https://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=72 Mihael Hategan <hategan(a)mcs.anl.gov> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |hategan(a)mcs.anl.gov Resolution| |EXPIRED --- Comment #25 from Mihael Hategan <hategan(a)mcs.anl.gov> 2013-02-16 22:48:10 --- I'm going to close the MolDyn bugs. I haven't heard of any efforts in this direction in the past few years. -- Configure bugmail: https://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

1 0

[Bug 933] New: Make the slurm --exclusive option configurable
by bugzilla-daemon＠mcs.anl.gov 15 Feb '13

15 Feb '13

https://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=933 Summary: Make the slurm --exclusive option configurable Product: Swift Version: trunk Platform: PC OS/Version: Mac OS Status: ASSIGNED Severity: normal Priority: P2 Component: Providers AssignedTo: davidk(a)ci.uchicago.edu ReportedBy: davidk(a)ci.uchicago.edu CC: hategan(a)mcs.anl.gov, davidk(a)ci.uchicago.edu, wozniak(a)mcs.anl.gov Need to make the behavior of --exclusive controllable by the user. You *may* be able to do this already using key="exclusive">false, but needs tested to make sure it works as expected. -- Configure bugmail: https://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

1 1

[Bug 889] New: Adjust local coasters for condor provider
by bugzilla-daemon＠mcs.anl.gov 15 Feb '13

15 Feb '13

https://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=889 Summary: Adjust local coasters for condor provider Product: Swift Version: trunk Platform: PC OS/Version: Mac OS Status: ASSIGNED Severity: normal Priority: P2 Component: Providers AssignedTo: davidk(a)ci.uchicago.edu ReportedBy: wilde(a)mcs.anl.gov CC: hategan(a)mcs.anl.gov, davidk(a)ci.uchicago.edu, wozniak(a)mcs.anl.gov A preliminary hack to make this work was contributed by Glen Hocky. David, can you see if you want to integrate and this this? If not, please re-assign to Mihael. WOuld be good to get this into 0.94. I *thought* it already worked, so Im glad that Glen has tried it and contributed an initial fix to the problem. Ive inserted below Glen's latest mail as well as his swift-support ticket. ---- Hey everyone, This is in reference to a problem I just sent to swift-support earlier. In order to get coasters to work on a condor system without a shared file system, I suggested putting worker.pl as an input file to condor which is then transferred. The problem is that this needs to be translated through from the coasters provider or via some environmental inputs that I don't know how to do However, here is a hacked proof of concept that "solves" the problem in a way specific to me. What it does is rather than write the randomly generated cscript*.pl file in the "arguments" string, the condor provider now skips that argument and throws in the hard coded path to worker.pl It would be nice for someone who actually knows what they're doing to make a general version of this fix Best Glen ------------------------ Index: provider-localscheduler/src/org/globus/cog/abstraction/impl/scheduler/condor/CondorExecutor.java =================================================================== --- provider-localscheduler/src/org/globus/cog/abstraction/impl/scheduler/condor/CondorExecutor.java (revision 3524) +++ provider-localscheduler/src/org/globus/cog/abstraction/impl/scheduler/condor/CondorExecutor.java (working copy) @@ -53,6 +53,16 @@ if ("MPI".equals(type)) { wr.write("universe = MPI\n"); } + else if ("parallel".equals(type)) { + wr.write("universe = parallel\n"); + } + else if ("local-cloud".equals(type)){ + wr.write("universe = vanilla\n"); + wr.write("should_transfer_files = YES\n"); + wr.write("when_to_transfer_output = ON_EXIT_OR_EVICT\n"); + wr.write("Transfer_Executable = true\n"); + wr.write("transfer_input_files = /mnt/hadoop/hockyg/swiftwork/worker.pl\n"); + } else if("grid".equals(type)) { grid = true; String gridResource = (String) spec.getAttribute("gridResource"); @@ -69,10 +79,15 @@ else { wr.write("universe = vanilla\n"); } + + //set account group if specified + writeAttr("AccountingGroup", "+AccountingGroup = ", wr); + if ("true".equals(spec.getAttribute("holdIsFailure"))) { wr.write("periodic_remove = JobStatus == 5\n"); } writeAttr("count", "machine_count = ", wr); + writeAttr("count", "request_cpus = ", wr); if (spec.getStdInput() != null) { wr.write("input = " + quote(spec.getStdInput()) + "\n"); } @@ -103,6 +118,9 @@ if (args != null && args.size() > 0) { wr.write("arguments = "); i = args.iterator(); + //note: space at end of next line is important + wr.write("/mnt/hadoop/hockyg/swiftwork/worker.pl "); + i.next(); while (i.hasNext()) { wr.write(quote((String) i.next())); if (i.hasNext()) { _______________________________________________ Swift-devel mailing list Swift-devel(a)ci.uchicago.edu https://lists.ci.uchicago.edu/cgi-bin/mailman/listinfo/swift-devel ---- Fri Nov 30 13:55:46 2012: Request 22371 was acted upon. Transaction: Ticket created by hockyg(a)uchicago.edu Queue: swift-support Subject: running on uc3 or local condor site Owner: Nobody Requestors: hockyg(a)uchicago.edu Status: new Ticket <URL: https://rt.ci.uchicago.edu/Ticket/Display.html?id=22371 > Hi Everyone, I'm trying to get my swift workflow running on UC3 again now that the system configuration has changed. For those who aren't familiar, it is now a condor pool where the only shared filesystem is a hadoop file system (HDFS), meaning that one is slow, not fully posix compatible, and apparently doesn't allow for setting files as executable. Anyway, I was wondering if anyone is running there yet or has solved the problems below. I will summarize some of the attempts I've made to get things to work and maybe you can offer some suggestions. Best, Glen p.s. By the way, I just updated my swift to whatever is newest in trunk as of an hour ago to make sure these aren't things that have been solved already --------------------------------------------------------------------------------------------- *** OPTION 1 - running with local condor provider *** First of all, to get anywhere with the new set up, I had to patch CondorExecutor.java and add > ... > else if ("local-cloud".equals(type)){ > wr.write("universe = vanilla\n"); > wr.write("should_transfer_files = YES\n"); > wr.write("when_to_transfer_output = > ON_EXIT_OR_EVICT\n"); > wr.write("Transfer_Executable = true\n"); > } > ... Otherwise there is an error b/c the stdin and stdout files can no longer be accessed Then in my sites file, I set among other things <profile namespace="globus" key="jobType">local-cloud</profile> <workdirectory>/mnt/hadoop/hockyg/swiftwork</workdirectory> and I do not set <scratch> such that everything writes to this directory The error I get in this case is > /mnt/hadoop/hockyg/swiftwork/glassRunFrozen-20121130-1335-yso4fbk4/shared/_swiftwrap: > line 402: 3: No such file or directory Which actually is there: > ls -l > /mnt/hadoop/hockyg/swiftwork/glassRunFrozen-20121130-1335-yso4fbk4/shared/_swiftwrap > -rw-rw-r-- 1 hockyg root 13067 Nov 30 13:35 > /mnt/hadoop/hockyg/swiftwork/glassRunFrozen-20121130-1335-yso4fbk4/shared/_swiftwrap But i suspect doesn't appear in time b/c HDFS is sow *** OPTION 2 - running with coasters *** This seems like a good option, b/c if coasters is running, it can work in a local $TMP directory that the head node can't see and do all the staging To do this, I switch my sites file to have <execution provider="coaster" url="localhost" jobManager="local:condor" /> Here, the problem is that the node cannot see ~/.globus/coasters/cscript* -> One solution would be to modify the condor_submit script to have this as a required input that is automatically transferred. I tried looking at the code for this and it seems like that might be difficult A solution I then tried (based on some useful information from Mike in an old listhost email ) was export SWIFT_USERHOME=/mnt/hadoop/hockyg Unfortunately, this *does not work*. It results in blank CondorSubmit scripts and cscript*.pl files, probably b/c HDFS does not support some I/O operation used by java to write these files *** OPTION 3 - running with persistent coasters *** (This method almost works, but I'd prefer the previous one) I tried to modify the instructions on http://www.ci.uchicago.edu/swift/guides/trunk/siteguide/siteguide.html i.e. setting export WORKER_LOCATION=/mnt/hadoop/hockyg/swiftwork export SHARED_FILESYSTEM=no and modifying start-coaster service to tell condor to transfer output files > *** > /home/hockyg/swift/devel/cog/modules/swift/dist/swift-svn/bin/start-coaster-service > 2012-11-30 12:41:51.880458030 -0600 > --- /home/hockyg/swift/devel/cog/modules/swift/bin/start-coaster-service > 2012-11-30 12:34:11.063383770 -0600 > *************** > *** 284,293 **** > leave_in_queue = FALSE > machine_count = $CONDOR_MACHINE_COUNT > +AccountingGroup="group_friends.$USER" > - should_transfer_files = YES > - when_to_transfer_output = ON_EXIT_OR_EVICT > queue $CONDOR_WORKERS > - > EOF > > if [ -n "$WORKER_RELAY_HOST" ]; then > --- 284,290 ---- *Now this actually get's a worker.pl running successfully!!! But has the error *Cannot find template for persistent-coasters B/c of this, I don't have a sites or tc file to work with *** Option 4 - treat as a grid site *** This has all been with local coasters and the condor provider on uc3-sub It is also possible I assume to pretend this is a remote grid site The problem here, I believe, is that filesystem will have to be gridftp, and I don't know what the correct url for that is or if it's set up yet -- Configure bugmail: https://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

1 1

[Bug 192] displeasing stack trace when pwd is not writable
by bugzilla-daemon＠mcs.anl.gov 15 Feb '13

15 Feb '13

https://bugzilla.mcs.anl.gov/swift/show_bug.cgi?id=192 David Kelly <davidk(a)ci.uchicago.edu> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED CC| |davidk(a)ci.uchicago.edu Resolution| |DUPLICATE --- Comment #1 from David Kelly <davidk(a)ci.uchicago.edu> 2013-02-15 13:00:44 --- *** This bug has been marked as a duplicate of bug 766 *** -- Configure bugmail: https://bugzilla.mcs.anl.gov/swift/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are watching all bug changes.

1 0