[DRBD-user] oom-killer

Lars Ellenberg lars.ellenberg at linbit.com
Mon Jun 23 09:58:21 CEST 2008

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


On Mon, Jun 23, 2008 at 03:49:13PM +0800, PN wrote:
> Hi all,
>  
> Currently we are using DRBD 8.2 on RHEL 4.5. We have 5 drbd partitions and
> works with GFS and multipath. It works well in the past few months.
> Last week, we increased the number of partitons to 12 and do a full sync, OOM
> occurred and we need to reboot the whole cluster.
> The 2 storages have dual Xeon 3GHz CPUs and 2GB RAM, is it too few to do the
> full sync?
> Our target is to increased to 30 partitions, how much memory shall be enough
> for the operational need on the storage servers?


30 drbd, all syncing in parallel,
using max-buffers (in unit of page size) of ~4000 to 8000:
 8000 * 4k * 30 --> around one GB of buffers in flight
with drbd only.
if you cannot handle that gracefully, down the "max-buffers" settings,
and serialize the resync of those resources.

-- 
: Lars Ellenberg                           http://www.linbit.com :
: DRBD/HA support and consulting             sales at linbit.com :
: LINBIT Information Technologies GmbH      Tel +43-1-8178292-0  :
: Vivenotgasse 48, A-1120 Vienna/Europe     Fax +43-1-8178292-82 :
__
please don't Cc me, but send to list -- I'm subscribed



More information about the drbd-user mailing list