MAILTO=REPL_MAILTO MAILFROM=REPL_MAILFROM PATH=/p4/common/bin:/p4/REPL_INSTANCE/bin:/usr/local/bin:/usr/local/sbin:/usr/bin:/usr/sbin:/bin:/sbin: SHELL=/bin/bash # common tasks INSTANCE=REPL_INSTANCE # Verify # If the Edge is using lbr.replication=cache, comment this out. 5 5 * * 6 [ -e /p4/common/bin ] && /p4/common/bin/p4verify.sh ${INSTANCE} # Checkpoints # Run recreate_db_checkpoint.sh on the first Saturday afternoon in Jan. and July. 25 13 * 1,7 6 [ $(date +\%d) -le 07 ] && /p4/common/bin/recreate_db_checkpoint.sh ${INSTANCE} 25 3 * * * [ -e /p4/common/bin ] && /p4/common/bin/daily_checkpoint.sh ${INSTANCE} # Replica Status 0 8 * * * [ -e /p4/common/bin ] && /p4/common/bin/replica_status.sh ${INSTANCE} > /dev/null
# | Change | User | Description | Committed | |
---|---|---|---|---|---|
#8 | 22622 | C. Thomas Tyler |
Renamed crontab templates files to be template.* rather than *.template. This is an 'optimize for lazy' tweak, since if we use a prefix, then the actual crontab files named of the form crontab.$USER.${HOSTNAME%%.*} won't have command line completion conficts with the templates. Added a ReadMe.md file describing the suggested (and with HMS required) naming convention for crontab files. Added a command summary doc for the sdp_sync.sh script that manages crontab files in a tight ship deployment. The reference to this file from the ReadMe.md file won't work now with this change in the dev branch, but will work once promoted to main. |
||
#7 | 21327 | Russell C. Jackson (Rusty) | Update to have mkdirs put the correct instance name into the crontab files. | ||
#6 | 19958 | Russell C. Jackson (Rusty) | Correcting crontab again... | ||
#5 | 19876 | Russell C. Jackson (Rusty) |
Corrected crontab to only run on 1st Sat of Jan. and July. |
||
#4 | 19113 | Russell C. Jackson (Rusty) |
Changed name of daily_backup.sh to daily_checkpoint.sh Changed name of weekly_backup.sh to recreate_db_checkpoint.sh Updated crontabs with new names, and changed to run recreate_db_checkpoint on the 1st Sat. of Jan. and July. For most companies, this is a better practice than recreating weekly per discussion with Anton. Remove solaris crontab since Solaris is pretty much dead, and we don't test on it. Updated docs to reflect name changes, and did a little clean other other sections while I was in there. |
||
#3 | 19112 | Russell C. Jackson (Rusty) |
Turn off database reorg per best practice recommendation from Anton and Michael S. Stop running weekly_backup.sh automatically since a compact database actually causes a performance hit. Added note to occasionally run weekly_backup.sh to recapture free space. |
||
#2 | 18859 | Russell C. Jackson (Rusty) |
Created script to rotate the journal and replay it to the offline database. Updated the crontab to run the new script before the weekly. Updated the edge crontab to account for the time adjustment in the master crontab. |
||
#1 | 18618 | Russell C. Jackson (Rusty) | Added a crontab for the edge servers. |