Hi Honza,
Ok. So you just changed libqb without recompilation of corosync?
That's what I did.
whole cluster to freeze with 100% CPU usage. So except for the libqb
version all nodes are the same, they just run different resources, the
This is problem. I mean, if they would be really same and running only
different libqb, or different resources, victim would be easy to find.
Apache ones are fine after many days of running, the postgres nodes are
not as I said.
I would really like to fix this issue if possible, so I'm open to any
ideas or suggestions.
Let's focus with theory that victim is libqb (because corosync didn't
changed). So this is probably libqb bug = nothing you can change by
changing configuration, ...
Can you please try to compile libqb from git (there are some fixes)?
If this doesn't help, can you please try to recompile corosync (maybe
there is some ABI break in libqb)? Let's see if this helps.
I compiled both Corosync and libqb from GitHub, latest versions. The
problem persists, after 58 minutes uptime Corosync uses 30% of RAM
(around 1 GB). I really don't know what else to do.
regards,
Domonkos
_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss
