Red Hat Bugzilla – Bug 565618
condor_submit fsync()s UserLog for each job
Last modified: 2010-10-14 12:08:51 EDT
Description of problem: $ echo -e 'cmd=/bin/sleep\nargs=1m\nlog=test.log\nqueue 100\n' | strace condor_submit 2>&1 | grep ^fsync | wc -l 100 $ echo -e 'cmd=/bin/sleep\nargs=1m\nlog=test.log\nqueue 1000\n' | strace condor_submit 2>&1 | grep ^fsync | wc -l 1000 Version-Release number of selected component (if applicable): All through at least condor 7.4.3-0.1 How reproducible: 100% Steps to Reproduce: 1. (see description) Expected results: Considerably fewer fsyncs, just 1 would be best
commit 579dffc9e99b6b50cce342f5468b3b3b8c3a95fb Author: Dan Bradley <dan@> Date: Thu Feb 11 16:35:55 2010 -0600 Only fsync user log once when writing multiple submit events. Previously, 'queue 100' caused 100 calls to fsync.
With the patch from #c1 applied: $ echo -e 'cmd=/bin/sleep\nargs=1m\nlog=test.log\nqueue 100\n' | strace ./release_dir/bin/condor_submit 2>&1 | grep ^fsync | wc -l 1
slated for condor 7.4.3-0.2
Problem reproduced on: $ condor -v $CondorVersion: 7.4.1 Dec 11 2009 BuildID: RH-7.4.1-0.7.1.el5 PRE-RELEASE $ $CondorPlatform: X86_64-LINUX_RHEL5 $ $ echo -e 'cmd=/bin/sleep\nargs=1m\nlog=test.log\nqueue 100\n' | strace condor_submit 2>&1 | grep ^fsync | wc -l 100 $ echo -e 'cmd=/bin/sleep\nargs=1m\nlog=test.log\nqueue 1000\n' | strace condor_submit 2>&1 | grep ^fsync | wc -l 1000
$CondorVersion: 7.4.3 Mar 16 2010 BuildID: RH-7.4.3-0.5.el4 PRE-RELEASE $ $CondorPlatform: X86_64-LINUX_RHEL4 $ $ echo -e 'cmd=/bin/sleep\nargs=1m\nlog=test.log\nqueue 100\n' | strace condor_submit 2>&1 | grep ^fsync | wc -l 1 $CondorVersion: 7.4.3 Mar 16 2010 BuildID: RH-7.4.3-0.5.el4 PRE-RELEASE $ $CondorPlatform: I386-LINUX_RHEL4 $ $ echo -e 'cmd=/bin/sleep\nargs=1m\nlog=test.log\nqueue 100\n' | strace condor_submit 2>&1 | grep ^fsync | wc -l 1 $CondorVersion: 7.4.3 Mar 16 2010 BuildID: RH-7.4.3-0.5.el5 PRE-RELEASE $ $CondorPlatform: X86_64-LINUX_RHEL5 $ $ echo -e 'cmd=/bin/sleep\nargs=1m\nlog=test.log\nqueue 100\n' | strace condor_submit 2>&1 | grep ^fsync | wc -l 1 $CondorVersion: 7.4.3 Mar 16 2010 BuildID: RH-7.4.3-0.5.el5 PRE-RELEASE $ $CondorPlatform: I386-LINUX_RHEL5 $ $ echo -e 'cmd=/bin/sleep\nargs=1m\nlog=test.log\nqueue 100\n' | strace condor_submit 2>&1 | grep ^fsync | wc -l 1 Problem resolved in version 7.4.3-0.5 >>> VERIFIED
Technical note added. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. New Contents: Logging 'condor_submit' fsync()s resulted in the creation of a log for each executed job. With this update, all outputs are logged in a single file.
Technical note updated. If any revisions are required, please edit the "Technical Notes" field accordingly. All revisions will be proofread by the Engineering Content Services team. Diffed Contents: @@ -1 +1 @@ -Logging 'condor_submit' fsync()s resulted in the creation of a log for each executed job. With this update, all outputs are logged in a single file.+condor_submit fsync()d the UserLog for each job submitted. With this update, the UserLog is only fsync()d once.
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2010-0773.html