Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On 30 Jul 2014, at 14:39, Lars Ellenberg <lars.ellenberg at linbit.com> wrote: > On Tue, Jul 29, 2014 at 12:09:51PM +0200, Bart Coninckx wrote: >> Not sure how the output is supposed to look like, but what I have now >> seems readable and provides quite some info. Does anyone can give me >> pointers on what to look for? Where can one determine what is the >> offending process? > > You are using an SCST patched kernel (iscsi_put_page_callback). > Apparently that patch is not compatible with DRBD, > or simply "forgets" to also patch DRBD as appropriate. > > You later posted that "disable_sendpage" seems to "improve stability". > > It changes timings, and thus also "life times" of page reference counts > within the network stack. Since the issue seems to be with page > reference counting, that may well "appear" to help. > > -- > : Lars Ellenberg > : LINBIT | Your Way to High Availability > : DRBD/HA support and consulting http://www.linbit.com > > DRBD® and LINBIT® are registered trademarks of LINBIT, Austria. > __ > please don't Cc me, but send to list -- I'm subscribed > _______________________________________________ > drbd-user mailing list > drbd-user at lists.linbit.com > http://lists.linbit.com/mailman/listinfo/drbd-user OK, thank you - I will ask on the SCST mailing list if a) the patch is really necessary and improves things considerably b) if they in the process of creating this patch think about what their code does do the DRBD module. I gather from your comments that the current fix (I still need some more days of testing before drawing any definitive conclusions) is not really solid, probably because it is a bit "hit and miss". Thank you for looking at the issue, Rgds, BC