[Drbd-dev] [drbdmanage-proxmox] cannot create/restore lxc container on non-leader nodes

[Yannis Milios]

Hello,

Sorry if this is not the right ML to report this, but thought it might good
to report it just in case that you are not already aware of it.

I'm using a 3 node Proxmox cluster backed by DRBD9 over ZFS.
All seem to work pretty well, apart from an issue with LXC containers.

The facts:
====================
cat /proc/drbd
version: 9.0.12-1 (api:2/proto:86-112)
GIT-hash: 7eb4aef4abbfba8ebb1afbcc30574df74db0063e build by root at pve2,
2018-03-14 08:23:43
Transports (api:16): tcp (9.0.12-1)

drbdmanage --version
drbdmanage 0.99.16; GIT-hash: a51c4e7858da69c4f56fc5b945c0ef586aae318b

drbdmanage-proxmox        2.1-1

drbd-utils                9.3.0-1

pveversion
pve-manager/5.1-46/ae8241d4 (running kernel: 4.13.13-6-pve
======================


The issue:
====================

When creating a new LXC container (or restoring from a LXC backup) on a
*non-leader* DRBD9 node, the process is failing with the following error:



> pong
> Operation completed successfully
> Operation completed successfully
> pong
> Operation completed successfully
> Operation completed successfully
> pong
> Operation completed successfully
> pong
> Operation completed successfully
> Operation completed successfully
> Waiting for server: ..
> mke2fs 1.43.4 (31-Jan-2017)
> Discarding device blocks:    4096/3932160 528384/3932160 1052672/3932160
> 1576960/3932160 2101248/3932160 2625536/3932160 3149824/3932160
> 3674112/3932160                 done
> Creating filesystem with 3932160 4k blocks and 983040 inodes
> Filesystem UUID: e2255445-4aca-4bfc-b130-c0f2d1324b17
> Superblock backups stored on blocks:
> 32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208
> Allocating group tables:   0/120         done
> Writing inode tables:   0/120         done
> Creating journal (16384 blocks): done
> Multiple mount protection is enabled with update interval 5 seconds.
> Writing superblocks and filesystem accounting information:   0/120
> done
> pong
> Operation completed successfully
> Operation completed successfully
> pong
> Operation completed successfully
> Operation completed successfully
> pong
> Operation completed successfully
> pong
> Operation completed successfully
> Operation completed successfully
> Waiting for server: .
> mke2fs 1.43.4 (31-Jan-2017)
> The file /dev/drbd/by-res/vm-100-disk-2/0 does not exist and no size was
> specified.
> pong
> Operation completed successfully
> Operation completed successfully
> TASK ERROR: command 'mkfs.ext4 -O mmp -E 'root_owner=0:0'
> /dev/drbd/by-res/vm-100-disk-2/0' failed: exit code 1


When executing the same task on a leader node, everything works perfectly.

I remember this same issue when creating qemu vms in the past, but
eventually it was fixed. Would it be possible to do the same for LXCs
please?

Regards,
Yannis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-dev/attachments/20180406/5adba053/attachment.htm>