After discussing w/ Krishnan, it happened to be an issue with the service's stop method. The stop method doesn't stop all the running processes so when it returns, the contract isn't empty and svc.startd waits until the stop timeout is reached to kill the remaining processes and place the service into maintenance.
-- Renaud Krishnan.Chandrashekar at Sun.COM wrote: > Hi Everyone, > > I am working on Sun Management center 4.0(SunMC 4.0). This uses postgres. > > And below is what happening... > > bash-3.00# ps -aef | grep post > root 20227 7453 0 21:01:26 pts/5 0:00 grep post > smcdbu 6914 6910 0 20:29:14 ? 0:00 /usr/bin/postmaster -D > /var/opt/SUNWsymon/db/data/SunMC > smcdbu 6910 1 0 20:29:13 ? 0:01 /usr/bin/postmaster -D > /var/opt/SUNWsymon/db/data/SunMC > smcdbu 11480 6910 0 20:39:27 ? 0:00 /usr/bin/postmaster -D > /var/opt/SUNWsymon/db/data/SunMC > smcdbu 11479 6910 0 20:39:27 ? 0:00 /usr/bin/postmaster -D > /var/opt/SUNWsymon/db/data/SunMC > smcdbu 8637 6910 0 20:29:35 ? 0:00 /usr/bin/postmaster -D > /var/opt/SUNWsymon/db/data/SunMC > smcdbu 6913 6910 0 20:29:14 ? 0:01 /usr/bin/postmaster -D > /var/opt/SUNWsymon/db/data/SunMC > smcdbu 6916 6914 0 20:29:14 ? 0:00 /usr/bin/postmaster -D > /var/opt/SUNWsymon/db/data/SunMC > > If I do, kill -9 6910 (parent 1) , all other children of 6910 become > zombies and become children of 1 and eventually all these will get > killed but service ends up in maintenance state. I was expecting SMF to > bring it back to a good state. This did not happen and SMF gave up with > a time out. > > Looking at your post, I felt this is what exactly the problem you also > faced. Could you please let me know how you went further about the issue. > > Is it an issue with SMF or the application? > Is there any value that I can tweak to get this working? > > Thanks in advance, > Krishnan > _______________________________________________ > smf-discuss mailing list > smf-discuss at opensolaris.org
