[Users] Error when starting ploop CT

[Roman Haefeli]

Hi all

I'm experiencing sporadic errors when starting CTs having their FS on a
ploop device. The thing is I can't reproduce the error when starting the
CTs manually, I only get them when the cluster managment (pacemaker)
starts them. 

The setup is a 3-node (hostnodes) cluster with pacemaker controlling the
CTs and their resources. The CT's ploop devices are hosted an NFS share
mounted on all nodes.

I am not sure there is something wrong with OpenVZ. I suspect more
something along the lines of the cluster starting the CTs too early,
before the NFS mount is fully ready or something. First, I thought the
error means that the root directory is not accessible/not existing at
start time. I changed the start script so that it will only try to start
the CT after making sure the CT's root directory is accessible. However,
I still get the errors sporadically.
Altough, it might not be an OpenVZ problem, the community still might be
help identify the issue. What does below error exactly mean?

Apr  8 12:04:12 virtuetest1 ManageVE(ploop9)[663341]: ERROR: Starting container... Adding delta dev=/dev/ploop37382 img=/virtual/vz/private/59/root.hdd/root.hdd (rw) Error in reread_part (ploop.c:988): BLKRRPART /dev/ploop37382: Input/output error Error in ploop_mount_fs (ploop.c:1017): Can't mount file system dev=/dev/ploop37382p1 target=/virtual/vz/root/59: No such device or address Failed to mount image: Error in ploop_mount_fs (ploop.c:1017): Can't mount file system dev=/dev/ploop37382p1 target=/virtual/vz/root/59: No such device or address [21]

This is on:
Debian 6.0.7
pacemaker 1.1.7
vzctl 4.2
vzkernel-2.6.32-042stab075.2

Any help appreciated.

Roman