[DRBD-user] What causes DRBD to stop listening on TCP ports ?

Ashish Shukla wahjava.ml at gmail.com
Tue Jan 8 10:20:42 CET 2008

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


I'm having a 2-node cluster, running DRBD 8.0.6 on CentOS 5 (amd64)
with Heartbeat 2.1.2. A gigabit interface is dedicated for DRBD on
both servers. Both server's interfaces (used by DRBD) are connected
via a cross-cable.

Today morning, when I started the servers, the primary server "A" is
not started. So, the secondary server "B" took over the resources,
became the DRBD primary node, and started serving clients. After few
hours, after fixing problem in server "A", when I started server "A"
again, it is not able to connect to DRBD running on server "B". I
restarted server "A" again, but same problem. I checked server "B" and
all my DRBD filesystems were mounted fine, and all of my DRBD
resources are in "Primary/Unknown" state. I tried to telnet to DRBD
ports of my server "B", but I get "Connection refused". So, to fix
this, I unmounted my DRBD filesystems, restarted DRBD and now DRBD
started listening on ports. And now due to this, split brain occured,
and I need to resync server "A" with server "B".

I've pasted the output of "cat /var/log/messages |fgrep kernel" on
server "B" at "http://pastebin.ca/raw/846436". The DRBD is using
interface "eth1".

Can anyone figure out from the above stuff, what could've caused DRBD
to stop listening on its TCP ports, hmm...? Is it due to change in
status of "eth1" interface, hmm...?

Ashish Shukla आशीष शुक्ल                      http://wahjava.wordpress.com/
·-- ·- ···· ·--- ·- ···- ·- ·--·-· --· -- ·- ·· ·-·· ·-·-·- -·-· --- --

More information about the drbd-user mailing list