[DRBD-user] DRBD crash and split brain

Ross Anderson rosander at dsotm.net
Mon Sep 14 00:35:19 CEST 2015

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


Storage system running areca raid with DRBD replication between two 
servers.

Scst provides FC and iscsi targets. I have a number of these systems in 
use however one in particular is giving me grief. From time to time a 
kernel panic is issued and DRBD gets into a very odd state. 6 DRBD 
devices replicated.
On the node with the kernel panic it continues to stay Primary/Secondary 
with no OOS numbers increasing. On secondary any number of devices will 
go WFConn while others stay up and continue to stay in sync. Spinlock 
with IRQsave appears. I've found some stability from sendpage disabled 
however that seems out of place since this is a baremetal device.

Any insight would be appreciated.

kernel 3.14.51
Gcc 4.8.5
drbd 8.4.6 git pull

http://pastebin.com/DSxj4WMi


Ross



More information about the drbd-user mailing list