OpenVZ Forum


Home » General » Support » IBM RAID controller re-mount lock (Using a RAID controller causes a system lock when doing a restore after a mount)
IBM RAID controller re-mount lock [message #36976] Wed, 05 August 2009 14:35 Go to previous message
maximilian.csuk is currently offline  maximilian.csuk
Messages: 2
Registered: August 2009
Junior Member
Dear OpenVZ community,

We have a system from IBM (System x3400) with an IBM ServeRAID 8K-I controller on which we are running Debian-Lenny 4.1.2-25. The kernel is openvz-kernel 2.6.26-2-openvz-686.

The containers are located on a separate ext3-partition and are mounted like this:
/dev/drbd0 on /mnt/hq-drbd-stor01 type ext3 (rw)

In particular, we experience a lockup when we try to un-mount and remount the container's partition.

The following steps work:
vzctl chkpnt 77
vzctl stop 77
/etc/init.d/openvz stop
/etc/init.d/openvz start
vzctl restore 77


But these don't work:
vzctl chkpnt 77
vzctl stop 77
/etc/init.d/openvz stop
umount /dev/drbd0
mount /dev/drbd0
/etc/init.d/openvz start
vzctl restore 77

The restore produces the lockup.

The output until it hangs:
Restoring VE ...
Starting VE ...
VE is mounted
        undump...
Adding IP address(es): 10.20.60.20
Setting CPU units: 1000
Configure meminfo: 65536

Nothing more is displayed after this.
The host shows 100% CPU (system-)usage, load goes up to about 10. From our understanding, a container's processes run inside a kernel-process on the host. The container's currently active process (as far as we can tell, it's a different one on any try) takes all the CPU from the host.

OpenVZ-version is 036test001.

We already tried changing the RAID-controller driver in use from aacraid-1.1.5.2449 to aacraid-1.1.5.2467 but without any luck.
Then we tried using USB-disks instead of the dev/drbd0 where the restore worked even after re-mounting, so the issue must lie within the RAID controller.
We also tried to run a DRBD device on two USB-disks instead of using ServeRAID attached disks this constellation also works, so we can assume that DRDB is working correctly

Is this a known issue? If yes, (how) can it be solved?
Is there a way to debug this and provide you and us with more information?

Thank you for your time.,
Maximilian Csuk
ipcenter.at
 
Read Message
Read Message
Read Message
Read Message
Previous Topic: Assigning an IP address to venet0 on the HN
Next Topic: Help Firefox
Goto Forum:
  


Current Time: Fri Aug 30 12:28:39 GMT 2024

Total time taken to generate the page: 0.03336 seconds