Description of problem: Condor does not start up properly and sends nagging mails 'obituaries'. Version-Release number of selected component (if applicable): condor-7.5.5-2.fc15 How reproducible: Always Steps to Reproduce: 1. yum install condor 2. service condor start 3. reboot Actual results: Nag mails, and errors in /var/log/condor/Masterlog (and others) similar to: 06/07/11 15:52:36 DaemonCore: ERROR: Can't open pid file /var/run/condor/condor_master.pid Expected results: No nagging. Additional info: Reason is that condor fails to create the directory /var/run/condor. It used to install it through its rpm, but now /var/run is on tmpfs, so its content disappears after every boot. Suggested fix: Add a file condor.conf in /etc/tmpfiles.d with contents: -- # Condor needs directory in /var/run d /var/run/condor 0775 condor condor --
Brian reports he has a patch for this (and systemd).
Confirmed on - $CondorVersion: 7.7.0 Jun 07 2011 PRE-RELEASE-UWCS $ $CondorPlatform: X86_64-Fedora_15 $ Only present after a system reboot. Currently seeing, MasterLog - 06/07/11 13:01:02 DaemonCore: ERROR: Can't open pid file /var/run/condor/condor_master.pid This is not a fatal issue, just an indication of a problem. The fatal issue is in the SchedLog - 06/07/11 13:02:20 (pid:1160) error opening watchdog pipe /var/run/condor/procd_pipe.SCHEDD.watchdog: No such file or directory (2) 06/07/11 13:02:20 (pid:1160) ProcFamilyClient: error initializing LocalClient 06/07/11 13:02:20 (pid:1160) ProcFamilyProxy: error initializing ProcFamilyClient 06/07/11 13:02:20 (pid:1160) ERROR "ProcD has failed" at line 587 in file /root/rpmbuild/BUILD/condor-7.6.0/src/condor_utils/proc_family_proxy.cpp -- # rpm -qV condor .......M. /var/lib/condor/execute missing /var/run/condor - Immediate workaround for missing /var/run/condor is to mkdir+chown condor.condor. Modification to execute is from drwxrwxrwt to drwxr-xr-x and does not appear to hinder execution of jobs.
Notes - 0) execute dir perms need validation 1) /tmp/condorLocks is being used instead of /var/lock/condor/local
Confirmed tmpfiles.d fix, # ls -al /etc/tmpfiles.d/condor.conf -rw-r--r--. 1 root root 74 Jun 7 13:15 /etc/tmpfiles.d/condor.conf # cat /etc/tmpfiles.d/condor.conf # Condor needs directory in /var/run d /var/run/condor 0775 condor condor
condor-7.7.0-0.3.fc15 has been submitted as an update for Fedora 15. https://admin.fedoraproject.org/updates/condor-7.7.0-0.3.fc15
condor-7.7.0-0.3.fc15 does NOT solve the problem, as the file /etc/tmpfiles.d/condor.conf is not what I suggested or what is in comment #4, but a kind of local install/test script. Also, I think it should not carry the file /var/lib/condor/condor_config.local as that file is replaced by the /etc/condor/config.d/* entries.
Typo in SOURCE2 (looked like SOURCE1). I'll also remove condor_config.local.
About condor_config.local: note that if it doesn't exist, then you need to throw a switch in the condor_config: REQUIRE_LOCAL_CONFIG_FILE = FALSE Otherwise it will choke on startup.
LOCAL_CONFIG_FILE is not defined, eliminating the need for REQUIRE_LOCAL_CONFIG_FILE=FALSE
condor-7.7.0-0.4.fc15 has been submitted as an update for Fedora 15. https://admin.fedoraproject.org/updates/condor-7.7.0-0.4.fc15
Package condor-7.7.0-0.4.fc15: * should fix your issue, * was pushed to the Fedora 15 testing repository, * should be available at your local mirror within two days. Update it with: # su -c 'yum update --enablerepo=updates-testing condor-7.7.0-0.4.fc15' as soon as you are able to. Please go to the following url: https://admin.fedoraproject.org/updates/condor-7.7.0-0.4.fc15 then log in and leave karma (feedback).
condor-7.7.0-0.5.fc15 has been submitted as an update for Fedora 15. https://admin.fedoraproject.org/updates/condor-7.7.0-0.5.fc15