http://old.nabble.com/file/p31742754/anomoly2011-05-30.rtf anomoly2011-05-30.rtf
Every now and then cron runs multiple times on Fedora 14. This is a new server and very little is running on it, only my colleageue and myself have jobs running. This does not happen all the time which makes it even harder to figure out why.
On 05/31/2011 07:41 PM, lhp1 wrote:
http://old.nabble.com/file/p31742754/anomoly2011-05-30.rtf anomoly2011-05-30.rtf
Every now and then cron runs multiple times on Fedora 14. This is a new server and very little is running on it, only my colleageue and myself have jobs running. This does not happen all the time which makes it even harder to figure out why.
The duplicate entries look like sub-processes giving the impression of it running twice, e.g.
May 30 05:00:10 flightdb crontab[1254]: (mysql) LIST (mysql) May 30 05:00:10 flightdb crontab[1253]: (mysql) LIST (mysql)
May 30 05:00:10 flightdb crontab[1255]: (mysql) REPLACE (mysql) May 30 05:00:10 flightdb crontab[1256]: (mysql) REPLACE (mysql)
This is supported by the sequence of the pid, indicating that the LIST and REPLACE are sub processes.
Is the TIME on the machine correctly sync'd via NTP ?
Albert.
p.s. make the effort to put your supporting data in the body of your email rather than in an RTF file :(, people really don't want to be looking at external and potentially dangerous attachments.
The server is synced to NTP. Last night another daily job called updatePhysicalView.sh executed 3 times. I've verified that there is only one instance of crond running. Is it possible that crond could be spawning other cron processes that cause the jobs to execute? It would be useful if there was a way to include the id of the process that executed the cron job. Is it possible to increase log levels in Fedora to get more info?
Jun 1 01:00:01 flightdb CROND[18810]: (root) CMD (/usr/lib64/sa/sa1 -S DISK 1 1) Jun 1 01:01:01 flightdb CROND[18842]: (root) CMD (run-parts /etc/cron.hourly) Jun 1 01:01:01 flightdb run-parts(/etc/cron.hourly)[18842]: starting 0anacron Jun 1 01:01:01 flightdb anacron[18852]: Anacron started on 2011-06-01 Jun 1 01:01:01 flightdb run-parts(/etc/cron.hourly)[18854]: finished 0anacron Jun 1 01:01:01 flightdb run-parts(/etc/cron.hourly)[18842]: starting mcelog.cron Jun 1 01:01:01 flightdb anacron[18852]: Normal exit (0 jobs run) Jun 1 01:01:01 flightdb run-parts(/etc/cron.hourly)[18861]: finished mcelog.cron Jun 1 01:10:01 flightdb CROND[19028]: (root) CMD (/usr/lib64/sa/sa1 -S DISK 1 1) Jun 1 01:12:01 flightdb CROND[19067]: (rfdb) CMD (/home/rfdb/prod/scripts/shell/updateTimePeriods.sh) Jun 1 01:15:01 flightdb CROND[19161]: (rfdb) CMD (/home/rfdb/prod/scripts/shell/updateDailyJobs.sh) Jun 1 01:18:02 flightdb CROND[19212]: (rfdb) CMD (/home/rfdb/prod/scripts/shell/updatePhysicalView.sh) Jun 1 01:18:12 flightdb CROND[19239]: (rfdb) CMD (/home/rfdb/prod/scripts/shell/updatePhysicalView.sh) Jun 1 01:18:22 flightdb CROND[19266]: (rfdb) CMD (/home/rfdb/prod/scripts/shell/updatePhysicalView.sh) Jun 1 01:20:01 flightdb CROND[19336]: (root) CMD (/usr/lib64/sa/sa1 -S DISK 1 1) Jun 1 01:30:01 flightdb CROND[19525]: (root) CMD (/usr/lib64/sa/sa1 -S DISK 1 1) Jun 1 01:40:01 flightdb CROND[19735]: (root) CMD (/usr/lib64/sa/sa1 -S DISK 1 1) Jun 1 01:50:01 flightdb CROND[19889]: (root) CMD (/usr/lib64/sa/sa1 -S DISK 1 1) Jun 1 02:00:01 flightdb CROND[20077]: (root) CMD (/usr/lib64/sa/sa1 -S DISK 1 1) Jun 1 02:01:01 flightdb CROND[20087]: (root) CMD (run-parts /etc/cron.hourly) Jun 1 02:01:01 flightdb run-parts(/etc/cron.hourly)[20087]: starting 0anacron Jun 1 02:01:01 flightdb anacron[20098]: Anacron started on 2011-06-01 Jun 1 02:01:01 flightdb run-parts(/etc/cron.hourly)[20100]: finished 0anacron Jun 1 02:01:01 flightdb run-parts(/etc/cron.hourly)[20087]: starting mcelog.cron
Root cause not known yet. However, the problem is directly related to the NTP service. The box is also running vmware under Fedora 14. Do not have any further info as I'm not a primary admin of the vmware. I do know that the cron issue was resolved when NTP service was stopped. Also, I found that the ntpdate service had no effect on syncing up the date when I did try to run it.
-
agraham -
lhp1