Re: Stability problems with ssi-1.9.1+drbd on debian

Markku Tavasti <[email protected]> writes:

> Ok, I changed drbd to use max 500k/s speed for sync. This definitely
> helps on issue, but does not solve it. Bit more load on cluster, and
> node2 crashes. 

I think the problem is clocks. Under vmware-server, clock doesn't
drift, it run's away. Like this:

ssi1:~#  for i in `seq 1 20`; do ntpdate ntp.xxx.xxx ; sleep 2; done
22 Sep 18:28:49 ntpdate[68841]: step time server i.p.n.o offset 4.921627 sec
22 Sep 18:28:58 ntpdate[68843]: step time server i.p.n.o offset 6.340782 sec
22 Sep 18:29:06 ntpdate[68845]: step time server i.p.n.o offset 6.283218 sec
22 Sep 18:29:15 ntpdate[68847]: step time server i.p.n.o offset 6.329403 sec
22 Sep 18:29:23 ntpdate[68849]: step time server i.p.n.o offset 6.269597 sec
22 Sep 18:29:32 ntpdate[68851]: step time server i.p.n.o offset 6.284451 sec
22 Sep 18:29:41 ntpdate[68853]: step time server i.p.n.o offset 6.338930 sec
22 Sep 18:29:49 ntpdate[68855]: step time server i.p.n.o offset 6.186271 sec
22 Sep 18:29:58 ntpdate[68857]: step time server i.p.n.o offset 6.349787 sec

Installing vmware-tools should fix this, but that's not
possible. After inserting vmware modules to ssi-kernel panics. 

-- 
M. Tavasti /  [email protected]  /   +358-40-5078254

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV