[DRBD-user] Page allocation failure (IPOIB, Infiniband, connected mode)
chibi at gol.com
Mon Mar 19 09:04:44 CET 2012
On Fri, 16 Mar 2012 13:55:17 +0100 Florian Haas wrote:
> On Wed, Mar 14, 2012 at 7:48 AM, Christian Balzer <chibi at gol.com> wrote:
> > Hello,
> > This is basically a repeat of:
> > http://lists.linbit.com/pipermail/drbd-user/2011-August/016758.html
> > 32GB RAM, Debian Squeeze, 3.2 (debian backport) kernel, 8.3.12 DRBD,
> > IPOIB in connected mode with a 64k MTU. Just 2 DRBD resources.
> > After encountering this for the first time (never showed up in two
> > weeks of stress testing, which only goes to prove that real life just
> > can't be simulated) I found the above article and changed the
> > following sysctls:
> > vm/min_free_kbytes = 262144
> > Lars hinted at "atomic reserves" in his reply, which particular
> > parameters are we talking about here?
> I had hoped for Lars to pitch in here, but I guess I'll give it a go
> instead. Note I'm certainly no kernel memory management expert, but
> I'm not aware of anything that would fit that description other than
> the vm.min_free_kbytes sysctl you've already mentioned.
Yeah, that was my assumption, too.
> SUSE's kernel documentation team, btw, lists these "page allocation
> failure" warnings as no cause for concern as long as they happen
Once or twice per day would fit that bill, however they still make me
I doubled the vm.min_free_kbytes again to 512MB and still got them at
times with particular high activity. Not sure if upping to 1GB would
actually make it go away, as reported free memory was several GB at least
once when such a failure was logged.
I guess I'll just keep an eye on it, these boxes are at about 30% of their
expected load/capacity (I/O, not space) now...
Christian Balzer Network/Systems Engineer
chibi at gol.com Global OnLine Japan/Fusion Communications
More information about the drbd-user