Opened 16 years ago
Last modified 12 years ago
#50 new enhancement
Install automatic cron failover
| Reported by: | andersk | Owned by: | |
|---|---|---|---|
| Priority: | major | Milestone: | |
| Component: | cron | Keywords: | |
| Cc: |
Description (last modified by quentin)
We need to deploy an automatic cron failover system, potentially based on quentin's system.
Change History (6)
comment:1 Changed 16 years ago by andersk
- Component changed from web to cron
comment:2 in reply to: ↑ description Changed 16 years ago by quentin
- Description modified (diff)
comment:3 Changed 14 years ago by mitchb
- Owner set to gdb
comment:5 Changed 13 years ago by adehnert
I gather (one of?) the major concerns with this were how you ensure that you don't have two servers running cron jobs at once, without allowing the servers to shut each other down and thus spike our HTTP load (ie, the STONITH issue).
Sticking cron in a cgroup and using that to kill all descendants, without needing to actually reboot the machine, is one option for this that has come up:
- -c scripts -i hacron, December 3rd, 2010
- -c scripts -i hacron, October 28th, 2010
comment:6 Changed 12 years ago by ezyang
- Type changed from defect to enhancement
Note: See
TracTickets for help on using
tickets.
We seem to be proceeding apace with Greg adapting Quentin's stuff into a pacemaker cluster based solution. I've backported the pacemaker-related packages to F11 and Greg has checked in a first pass at an OCF resource for hacron.