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Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!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Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!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#31597 - [postgresql] Use of Timeout(Start)Sec in postgresql.service
Attached to Project:
Arch Linux
Opened by Marti (intgr) - Tuesday, 18 September 2012, 10:40 GMT
Last edited by Dan McGee (toofishes) - Tuesday, 18 September 2012, 14:09 GMT
Opened by Marti (intgr) - Tuesday, 18 September 2012, 10:40 GMT
Last edited by Dan McGee (toofishes) - Tuesday, 18 September 2012, 14:09 GMT
|
DetailsDescription: As I understand, the systemd "TimeoutSec=120" value in postgresql.service means that the service will be killed if it doesn't start up in 2 minutes. This is potentially bad in configurations where checkpoint_segments/checkpoint_timeout have been increased. These settings are a compromise between crash-recovery time and runtime performance. When pushed to the maximum, startup recovery can take many minutes (as much as the time between checkpoints).
Do we care about this? Additional info: * package version(s) postgresql 9.2.0-1 |
This task depends upon
Closed by Dan McGee (toofishes)
Tuesday, 18 September 2012, 14:09 GMT
Reason for closing: Won't fix
Additional comments about closing: Systemd service file is overrideable by the end user, so it can be fixed there.
Tuesday, 18 September 2012, 14:09 GMT
Reason for closing: Won't fix
Additional comments about closing: Systemd service file is overrideable by the end user, so it can be fixed there.
Comment by Dan McGee (toofishes) -
Tuesday, 18 September 2012, 14:01 GMT
Please change the service file yourself to override this. I provided sane defaults, as most people aren't using Arch + PostgreSQL in a production environment. My biggest concern was making sure the pg_ctl timeout value matched the systemd timeout value.