[DRBD-user] Re: meta connection shut down by peer

Lars Ellenberg Lars.Ellenberg at linbit.com
Fri Sep 24 13:10:28 CEST 2004

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.

/ 2004-09-24 11:07:03 +0800
\ Seki Lau:
> Dear all,
> I read the post on 12 Aug
> http://marc.free.net.ph/message/20040812.104624.fbe30e06.html
> and I am now facing the same problem.
> I have machine  M0 and M1. M0 is startup. No problem at all. It
> becomes Primary and the states is Primary/unknown
> Then I strart M1. No problem. It becomes Secondary/Primary (I expect
> this result for every secondary).
> Then I power down M0. M1 becomes primary (expected result).
> Then I power up M0. Problem occurs here.
> M0: Secondary/unknow
> M1: Primary/unknow.
> I check the /var/log/messages in M0. it gives that
> meta connection shut down by peer.
> asender terminated
> sock was shut down by peer

it is not enough to look at the log of only *one* node,
the other node probably had a reason to do so...

> Then I power restart  M1, it becomes normal again:
> M0 Primary/Secondary
> M1 Secondary/Primary
> I am using drbd 0.7.4 and I find from mail list that, it is
> recommanded to start drbd before heartbeat(M0 and M1 do so).
> So, what should be the source of problem and how to fix it?

my guess is that your damn init system "forgets" about shutting down
drbd early enough, or at all. and this might be related to the drbd
init script not bothering with /var/lock/subsystem ...
well, we "fixed" that, so try again with recent svn:
 svn co http://svn.drbd.org/drbd/branches/drbd-0.7


	Lars Ellenberg

please use the "List-Reply" function of your email client.

More information about the drbd-user mailing list