Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
FS#44299 - [postgresql]Shutdown/startup oddity!
Attached to Project:
Arch Linux
Opened by Neil Darlow (neildarlow) - Sunday, 22 March 2015, 11:12 GMT
Last edited by Dan McGee (toofishes) - Sunday, 24 May 2015, 13:51 GMT
Opened by Neil Darlow (neildarlow) - Sunday, 22 March 2015, 11:12 GMT
Last edited by Dan McGee (toofishes) - Sunday, 24 May 2015, 13:51 GMT
|
DetailsDescription: I restarted my Arch system (x86_64) running Postgresql using "systemctl reboot". Following the boot the postgresql service was reported as being in a failed state by systemctl status -l. The stated reason was that a postmaster process may already be running at startup time.
Looking into the pg_log files I saw a message to the effect that Postgresql had not been properly shutdown and the following, manual, start initiated a recovery process. I see that the postgresql.service file has a TimeoutSec=120 and postgresql is started with a -w 120 option. Are these values adequate for a proper and complete shutdown/startup sequence? Additional info: postgresql 9.4.1-1 postgresql-libs 9.4.1-1 -- Reboot -- Mar 21 21:45:37 optimus postgres[556]: LOG: redirecting log output to logging collector process <-- Boot, postgresql.service enters failed state. Mar 21 21:45:37 optimus postgres[556]: HINT: Future log output will appear in directory "pg_log". Mar 21 21:45:41 optimus postgres[556]: pg_ctl: this data directory appears to be running a pre-existing postmaster Mar 21 21:45:41 optimus postgres[556]: pg_ctl: could not start server Mar 21 21:45:41 optimus postgres[556]: Examine the log output. Mar 21 21:57:40 optimus postgres[1786]: LOG: redirecting log output to logging collector process <-- Manual start of postgresql following failed state. Mar 21 21:57:40 optimus postgres[1786]: HINT: Future log output will appear in directory "pg_log". -- pg_log/postgresql-2015-03-21_214537.log -- LOG: database system was shut down at 2015-03-21 21:44:33 GMT LOG: database system is ready to accept connections LOG: autovacuum launcher started -- pg_log/postgresql-2015-03-21_215740.log -- LOG: database system was interrupted; last known up at 2015-03-21 21:45:37 GMT LOG: database system was not properly shut down; automatic recovery in progress LOG: record with zero length at 0/EC1DF68 LOG: redo is not required LOG: database system is ready to accept connections LOG: autovacuum launcher started Steps to reproduce: Reboot system and observe journalctl and pg_log output. |
This task depends upon
Comment by Dan McGee (toofishes) -
Sunday, 24 May 2015, 13:51 GMT
These values may not always be appropriate, especially if you have a large cluster where recovery takes a significant amount of time. However, in these cases, you should customize or override the unit configuration file yourself; we won't be able to cater to all audiences with our out of the box configuration file.