[DRBD-user] DRBD9 LVM2 Backed issue

cgasmith at comcast.net cgasmith at comcast.net
Mon Dec 14 23:14:32 CET 2015

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


I have drbd9 in a 3 node mesh, all seems normal (albeit 3 hours on a 10G network to initial sync 3Tb) . All nodes are backed by by an LVM allocated on top of a hardware raid, makeing the /dev/drbd1 on each node. For debugging, auto-promote is disabled by config. 

Each node has: 
/dev/vg_on_my_local_raid/local_for_drbd 

drbd uses above for its storage 
which serves as a pv for 

/dev/vg_reflected 

It seems that on the primary node that created /dev/vg_reflected (and test lv's inside it) has no problem to access 

vgchange /dev/vg_reflected -an 

primary node -> secondary 

Different node: 
drbd -> primary 
vgchange /dev/vg_reflected -ay 

results in IOCTL failures that wont mount. 

kernel: device-mapper: table: 253:4: linear: dm-linear: Device lookup failed 
kernel: device-mapper: ioctl: error adding target to table 


Upon several searches, it seems that this possibly occurs when LVM2 metadata is incorrect or corrupt. 
Though Its not corrupt as going back to the primary-creating-node works fine. 

So, I'm a bit confused where to look to solve the problem, it seems to me there is something that is node specific that should be blanked or not writing, I've tried crusted LVM. but that really doesnt do much except put vg control in ccs/cman, not of value for a single primary node configuration 1 of 3. 

Any suggestions would be appreciated, have considered wiping the whole node and create araid /dev/sdx device to use as the backing device, but that looses the flexibility on resize online adding pv's to the backing device. 




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20151214/00968300/attachment.htm>


More information about the drbd-user mailing list