[DRBD-user] OCFS2 over DRBDv8

Rene Mayrhofer rene.mayrhofer at gibraltar.at
Tue Feb 27 21:41:47 CET 2007

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Am Dienstag, 27. Februar 2007 19:09 schrieb lars:
> go, bug them!
I bugged the Debian kernel team instead - #412748.

> an other hint: you definetly want to increase the ocfs2 heartbeat
> "retries" so it is much higher than any drbd timeout might be.
> with the default of (iirc) 7, both ocfs2 nodes may decide to fence
> themselves whithin some seconds when the network has some hickups.
Is O2CB_HEARTBEAT_THRESHOLD=16 considered to be enough (i.e. 30s fencing time 
when I understand it correctly, compared to the drbd defaults of between 6 
and 10s)?

> and try to not run into oom conditions...
How would it if only a few user processes are running? Are the drbd and/or 
ocfs kernel threads known to leak memory? Or are you referring to something 
else here?

> > Or is GFS2 better or even the "recommended" file system for
> > primary/primary drbd volumes?
>
> we don't "recommend" anything yet.  I believe ocfs2 is more lightweight
> and would be easier to setup/maintain.  but that is just me...
But are there any practical experiences right now or am I trying 
something "yes, it might work, but is right now completely 
untested/unsupported/not expected to work/brkoen und you get to keep the 
pieces if your machine transforms into a fireball"?

> hope we hear a happy success story soon?
Yeah, me too...

Rene

-- 
-------------------------------------------------
Gibraltar firewall       http://www.gibraltar.at/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20070227/4d148d40/attachment.pgp>


More information about the drbd-user mailing list