Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
/ 2005-10-10 19:58:45 +0400 \ Eugene Crosser: > Lars Ellenberg wrote: > > >>I would like to hear from people: > >>- is there anybody who successfully use "protocol A" at all? > > > > just as a side note, as long as no failover was involved to reproduce > > the error, the drbd protocol is irrelevant. > > Protocol is indeed irrelevant to the error: we reproduced the problem > while protocol was 'C'. > > Today, we ran this test: filesystem was mounted on /dev/md3 (without > drbd), and we generated high traffic on the network interface (by > running netcat from /dev/zero on one host to /dev/null on another). > There where NO filesystem errors. > > We are planning to run a test when the filesystem is mounted from drbd, > and the gigabit interface is completely idle. When we do that, we'll > know for sure, but so far it *seems* that there is no relation between > the appearance of errors and network activity. > > Another test that we ran: we waited until a filesystem error on drbd > device appeared, did *not* run fsck, removed drbd module and mounted the > filesystem from md device. It works for almost 24 hours now, there are > no errors! Apparently, the filesystem errors indeed appear only in the > in-memory copy only. I figure that this is an important symptom to > narrow down the search. > > I'll come back when we have more information. thank you for investigating this systematically! kernel version, distribution, smp, preemt, memory, cpu, storage subsystem, ... ? -- : Lars Ellenberg Tel +43-1-8178292-0 : : LINBIT Information Technologies GmbH Fax +43-1-8178292-82 : : Schoenbrunner Str. 244, A-1120 Vienna/Europe http://www.linbit.com : __ please use the "List-Reply" function of your email client.