Red Hat Bugzilla – Bug 456482
submit -spool and transfer_executable = false
Last modified: 2010-10-14 11:57:31 EDT
Seems when you run condor_submit -spool on a job that has transfer_executable = false the executable gets transferred, at least to the spool directory. This seems like a bug to me.
As of at least 7.4.2-0.3 $ echo "cmd=/bin/sleep\nargs=1m\ntransfer_executable=false\nqueue 1" | condor_submit Submitting job(s). 1 job(s) submitted to cluster 1. $ echo "cmd=/bin/sleep\nargs=1m\ntransfer_executable=false\nqueue 1" | condor_submit -spool Submitting job(s). 1 job(s) submitted to cluster 2. $ echo "cmd=/bin/sleep\nargs=1m\ntransfer_executable=true\nshould_transfer_files=yes\nwhen_to_transfer_output=on_exit_or_evict\nqueue 1" | condor_submit -spool Submitting job(s). 1 job(s) submitted to cluster 4. $ echo "cmd=/bin/sleep\nargs=1m\ntransfer_executable=false\nshould_transfer_files=yes\nwhen_to_transfer_output=on_exit_or_evict\nqueue 1" | condor_submit -spool Submitting job(s). 1 job(s) submitted to cluster 5. $ echo "cmd=/bin/sleep\nargs=1m\ntransfer_executable=false\nshould_transfer_files=yes\nwhen_to_transfer_output=on_exit_or_evict\nqueue 1" | condor_submit -remote robin Submitting job(s). 1 job(s) submitted to cluster 6. $ echo "cmd=/bin/sleep\nargs=1m\ntransfer_executable=true\nshould_transfer_files=yes\nwhen_to_transfer_output=on_exit_or_evict\nqueue 1" | condor_submit -remote robin Submitting job(s). 1 job(s) submitted to cluster 7. $ l spool/cluster* spool/cluster2.proc0.subproc0: total 0 spool/cluster2.proc0.subproc0.tmp: total 0 spool/cluster4.proc0.subproc0: total 32K 32K -rwxr-xr-x. 1 matt matt 26K 2010-01-16 09:11 sleep* spool/cluster4.proc0.subproc0.tmp: total 0 spool/cluster5.proc0.subproc0: total 0 spool/cluster5.proc0.subproc0.tmp: total 0 spool/cluster6.proc0.subproc0: total 0 spool/cluster6.proc0.subproc0.tmp: total 0 spool/cluster7.proc0.subproc0: total 32K 32K -rwxr-xr-x. 1 matt matt 26K 2010-01-16 09:14 sleep* spool/cluster7.proc0.subproc0.tmp: total 0 This is good, only when transfer_executable=true does the executable get copied to the spool.
Testing on current stable release: $CondorVersion: 7.4.1 Dec 11 2009 BuildID: RH-7.4.1-0.7.1.el5 PRE-RELEASE $ $CondorPlatform: I386-LINUX_RHEL5 $ Submitted 2 jobs: [test@dell-pesc1425-01 ~]$ echo -e "cmd=/bin/sleep\nargs=1m\ntransfer_executable=false\nqueue 1" | condor_submit Submitting job(s). 1 job(s) submitted to cluster 10. -------- [test@dell-pesc1425-01 ~]$ echo -e "cmd=/bin/sleep\nargs=1m\ntransfer_executable=true\nshould_transfer_files=yes\nwhen_to_transfer_output=on_exit_or_evict\nqueue 1" | condor_submit -spool Submitting job(s). 1 job(s) submitted to cluster 11. Results: [root@dell-pesc1425-01 spool]# condor_q -run -- Submitter: dell-pesc1425-01.rhts.eng.bos.redhat.com : <10.16.64.52:38474> : dell-pesc1425-01.rhts.eng.bos.redhat.com ID OWNER SUBMITTED RUN_TIME HOST(S) 10.0 test 3/16 07:03 0+00:00:02 slot1@dell-pesc1425-01.rhts.eng.bos.redhat.com 11.0 test 3/16 07:04 0+00:00:02 slot2@dell-pesc1425-01.rhts.eng.bos.redhat.com -------- [root@dell-pesc1425-01 spool]# ls -la cluster* cluster11.proc0.subproc0: total 40 drwxr-xr-x 2 test test 4096 Mar 16 07:04 . drwxr-xr-x 9 condor condor 4096 Mar 16 07:04 .. -rwxr-xr-x 1 test test 19564 Mar 16 07:01 sleep cluster11.proc0.subproc0.tmp: total 16 drwxr-xr-x 2 test test 4096 Mar 16 07:04 . drwxr-xr-x 9 condor condor 4096 Mar 16 07:04 .. No such local copy of sleep found in spool when transfer_executable=false Same commands on other archs: $CondorVersion: 7.4.1 Dec 11 2009 BuildID: RH-7.4.1-0.7.1.el4 PRE-RELEASE $ $CondorPlatform: X86_64-LINUX_RHEL4 $ [root@dell-pe1855-02 spool]# ls -la cluster* cluster2.proc0.subproc0: total 44 drwxr-xr-x 2 condor condor 4096 Mar 16 07:09 . drwxr-xr-x 5 condor condor 4096 Mar 16 07:09 .. -rwxr-xr-x 1 condor condor 22040 Mar 16 07:06 sleep cluster2.proc0.subproc0.tmp: total 16 drwxr-xr-x 2 condor condor 4096 Mar 16 07:09 . drwxr-xr-x 5 condor condor 4096 Mar 16 07:09 .. [root@dell-pe1855-02 spool]# -------- $CondorVersion: 7.4.1 Dec 11 2009 BuildID: RH-7.4.1-0.7.1.el5 PRE-RELEASE $ $CondorPlatform: X86_64-LINUX_RHEL5 $ [root@nec-em11 spool]# ls -la cluster* && condor_q -run cluster2.proc0.subproc0: total 44 drwxr-xr-x 2 test test 4096 Mar 16 07:13 . drwxr-xr-x 5 condor condor 4096 Mar 16 07:13 .. -rwxr-xr-x 1 test test 21984 Mar 16 07:10 sleep cluster2.proc0.subproc0.tmp: total 16 drwxr-xr-x 2 test test 4096 Mar 16 07:13 . drwxr-xr-x 5 condor condor 4096 Mar 16 07:13 .. -------- $CondorVersion: 7.4.1 Dec 11 2009 BuildID: RH-7.4.1-0.7.1.el4 PRE-RELEASE $ $CondorPlatform: I386-LINUX_RHEL4 $ cluster2.proc0.subproc0: total 40 drwxr-xr-x 2 test users 4096 Mar 16 07:17 . drwxr-xr-x 5 condor condor 4096 Mar 16 07:18 .. -rwxr-xr-x 1 test users 17340 Mar 16 07:14 sleep cluster2.proc0.subproc0.tmp: total 16 drwxr-xr-x 2 test users 4096 Mar 16 07:17 . drwxr-xr-x 5 condor condor 4096 Mar 16 07:18 ..
Testing on candidate: Command used: echo -e "cmd=/bin/sleep\nargs=1m\ntransfer_executable=false\nqueue 1" | condor_submit && echo -e "cmd=/bin/sleep\nargs=1m\ntransfer_executable=true\nshould_transfer_files=yes\nwhen_to_transfer_output=on_exit_or_evict\nqueue 1" | condor_submit -spool -------- $CondorVersion: 7.4.3 Mar 16 2010 BuildID: RH-7.4.3-0.5.el4 PRE-RELEASE $ $CondorPlatform: I386-LINUX_RHEL4 $ cluster6.proc0.subproc0: total 40 drwxr-xr-x 2 test users 4096 Mar 17 12:50 . drwxr-xr-x 7 condor condor 4096 Mar 17 12:50 .. -rwxr-xr-x 1 test users 17340 Mar 17 12:47 sleep cluster6.proc0.subproc0.tmp: total 16 drwxr-xr-x 2 test users 4096 Mar 17 12:50 . drwxr-xr-x 7 condor condor 4096 Mar 17 12:50 .. -------- $CondorVersion: 7.4.3 Mar 16 2010 BuildID: RH-7.4.3-0.5.el4 PRE-RELEASE $ $CondorPlatform: X86_64-LINUX_RHEL4 $ cluster7.proc0.subproc0: total 44 drwxr-xr-x 2 condor condor 4096 Mar 17 12:28 . drwxr-xr-x 11 condor condor 4096 Mar 17 12:28 .. -rwxr-xr-x 1 condor condor 22040 Mar 17 12:25 sleep cluster7.proc0.subproc0.tmp: total 16 drwxr-xr-x 2 test users 4096 Mar 17 12:28 . drwxr-xr-x 11 condor condor 4096 Mar 17 12:28 .. -------- $CondorVersion: 7.4.3 Mar 16 2010 BuildID: RH-7.4.3-0.5.el5 PRE-RELEASE $ $CondorPlatform: I386-LINUX_RHEL5 $ cluster13.proc0.subproc0: total 40 drwxr-xr-x 2 test test 4096 Mar 17 12:36 . drwxr-xr-x 11 condor condor 4096 Mar 17 12:36 .. -rwxr-xr-x 1 test test 19564 Mar 17 12:33 sleep cluster13.proc0.subproc0.tmp: total 16 drwxr-xr-x 2 test test 4096 Mar 17 12:36 . drwxr-xr-x 11 condor condor 4096 Mar 17 12:36 .. -------- $CondorVersion: 7.4.3 Mar 16 2010 BuildID: RH-7.4.3-0.5.el5 PRE-RELEASE $ $CondorPlatform: X86_64-LINUX_RHEL5 $ cluster13.proc0.subproc0: total 44 drwxr-xr-x 2 test test 4096 Mar 17 12:47 . drwxr-xr-x 5 condor condor 4096 Mar 17 12:47 .. -rwxr-xr-x 1 test test 21984 Mar 17 12:44 sleep cluster13.proc0.subproc0.tmp: total 16 drwxr-xr-x 2 test test 4096 Mar 17 12:47 . drwxr-xr-x 5 condor condor 4096 Mar 17 12:47 .. No such regresion found in candidate. >>>> 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: Running the "condor_submit -spool" command on a job that has the "transfer_executable" variable set to 'false', the executable gets transferred to the spool directory. With this update, the executable is no longer transferred to the spool directory.
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