There are several cron jobs that help the various versions of the portal run smoothly. From Geoff Avila, regarding cron jobs to remove old workspace directories. ------------------------------------------------------ Under portal2prod (or test) in the /logs dir, remove anything with the pattern *.log.20* that's older than 3 0 days. Likewise, in the /logs/finished_logs dir, remove files that match that p attern that are older than 30 days. in sge_jobs and bwbatch_jobs, do the same for any files or directories that havn 't been touched in 30 days; recursively. Likewise, recursively remove anything m ore than a day old under /tmp/treedraw. The portal2test script is run by portal2test at midnight Sunday on 8ball. The po rtal2prod script is run at fifteen past midnight on snooker by the portal2prod u ser, and again five minutes later on snooker by the bwbatch user. Between those three, they should scan every file under the /archives/science/ngbt/portal* dirs. The two scripts themselves are in the portal2prod user's ~/scripts dir. ------------------------------------------------------ I checked those two scripts plus the others we use into sdk/scripts. So what we've got running is: ------------------------------------------------------------------------------------- portal2test@8ball midnight Sun portal2_test_exp.sh 10pm Sun update_crl.sh portal2prod@snooker 12:15 Sun portal2_prod_exp.sh bwbatch@snooker 12:15 Sun portal2_prod_exp.sh midnight, daily kill_long_jobs.sh portal2prod@qball 10pm, 1st of month update_crl.sh 10pm daily tg_ws_cleanup.sh 1am daily restart_tomcat.sh (too many open files issue) -------------------------------------------------------------------------------------