[DRBD-user] High iowait when other node is down

Lars Ellenberg lars.ellenberg at linbit.com
Tue Oct 2 15:26:54 CEST 2007

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, Oct 01, 2007 at 05:52:47PM +0200, Francesco Ciocchetti wrote:
> Stefan Seifert wrote:
> > Francesco Ciocchetti wrote:
> > > I'm experiencing a similar problem with 8.0.4 DRBD even if not so strong
> > > as Mark wrote. My system are still usable but my iowait is around 90% on
> > > a pseudo 400% cpu system (2 HT CPU that seems 4 to the system).
> >
> > Could this be the dead lock problem? DRBD < 8.0.6 is known not to work
> > on kernel 2.6.22.
> Actually i'm having this problem on a 2.6.18 kernel version ... is this
> bug known on this kernel too?

this "deadlock problem" on 2.6.22 is not a "bug".
it is an incompatibility with modifications in upstream kernel.

depending on what kernel you have, this incompatible modification may
already be present in a vendor kernel labeled 2.6.18.

> What kind of back effect can i expect?

"back effect" of the resulting deadlock is, well, deadlock.
no more io progress possible on drbd, reboot required.

if you still can write to drbd, though maybe slowly,
then what you see is something different.

: 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 use the "List-Reply" function of your email client.

More information about the drbd-user mailing list