Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
/ 2005-11-07 18:30:53 +0100 \ Markus Marquardt: > Hi, > > we encounter the same "freeze" problem with our dell PE2850 servers > (2x Xeon 3 GHz, 4 GB ram, Perc Raid 4/Di (Megaraid), latest FC4 kernel > 2.6.13, drbd-0.7.13, latest BIOS & raid firmware) as some other people > here. DRBD 0.7.13 contains a bug (introduced in 0.7.12, while fixing a much completely harmless other buglet), that triggers on SMP only, and will freeze (deadlock) a SyncSource Primary, if it is busy io-wise. this is fixed in 0.7.14. you should have received an announcement via the drbd-announce list. > We can reproduce this by: > > o On secondary: drbdadm invalidate drbd0 > o On primary: cat /dev/zero > /drbd/mybigfile exactly. > Another oddity after we've done some repartitioning of the underlying disk partitions: > > When stopping heartbeat on a cluster node (which implies a /etc/ha.d/resource.d/drbddisk stop) some messages like > > ... drbd_md_sync_page_io ... out of range ... > > appear on the console. What is this? well. that is an other issue. do you use "external" or "internal" drbd meta data? how exactly did you "repartiton"? can you provide cat /proc/partitions drbdadm dump drbdsetup /dev/drbd0 show # (or which device shows it) and the _complete_ line from syslog (i.e. including the ugly numbers) just the first and last line, maybe, in case there are many (probably with consecutive numbers) thanks, -- : 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.