OpenVZ Forum: Support » Unable to stop container

Home » General » Support » Unable to stop container (operation timed out)

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Unable to stop container [message #43574]

Tue, 27 September 2011 18:51

Rene
Messages: 40
Registered: September 2006

Member

I was messing with checkpointing ("vzctl chkpnt") when suddenly the VE just became unresponsive. It's a newly created test VE with just a couple of basic processes, but it's worrying that I can't do anything with it.

# vzctl restore 1401 --kill
Removing stale lock file /vz/lock/1401.lck
Error: No checkpointing support, unable to open /proc/rst: No such file or directory
# vzctl stop 1401
Stopping container ...
Unable to stop container: operation timed out
# vzctl enter 1401
enter into CT 1401 failed
# service vz restart
Shutting down CT 1401
Shutting down CT 1401
Shutting down CT 1401

I can start and stop other VE's no problem.

This is a Centos 6 server with OpenVZ 2.6.32-042stab037.1. Yes I saw that some people have problems with it, but the new UBC and vswap system makes it very very tempting to check it out.

Could this be related?

Would it be better to go with SL6.1? I kinda like the idea, after all it was a guy in CERN who invented the www. OTOH Centos has been around for years. I've been using Centos since version 3 and kinda feel a little attached to it.

Report message to a moderator

Re: Unable to stop container [message #43577 is a reply to message #43574]

Wed, 28 September 2011 07:12

Rene
Messages: 40
Registered: September 2006

Member

Not much activity on this forum huh...

Report message to a moderator

Re: Unable to stop container [message #43613 is a reply to message #43574]

Sat, 01 October 2011 13:25

Ales
Messages: 330
Registered: May 2009

Senior Member

Well, probably not many users have run into this problem yet. I know I haven't... Is your VM also CentOS 6, or..?

I'd start troubleshooting with:

1) check if vzcpt and vzrst kernel modules are loaded
2) do you use iptables? Even if not, do you have scripts /sbin/iptables-save and /sbin/iptables-save ? Checkpointing uses them
3) do you have any extra mounts within VM? Any bind mounts..?
4) do you use openvz quotas?
5) if all fails to produce any hints, I'd search trough bugzilla for anything related

BTW, the above kernel modules should be loaded by openvz, they are present on our servers & test machines. We use SL 6 and Fedora 14, but it should be exactly the same on CentOS 6.

Report message to a moderator

Re: Unable to stop container [message #43615 is a reply to message #43613]

Sun, 02 October 2011 08:46

Rene
Messages: 40
Registered: September 2006

Member

1) they're loaded on the hardware node if that's what you mean. Not in the container though.
2) /sbin/iptables-save is a binary executable. You asked /sbin/iptables-save or /sbin/iptables-save - same thing - typo?
3) The offending container has meanwhile been destroyed, but another container created at the same time shows
Filesystem Size Used Avail Use% Mounted on
/dev/simfs 20G 3.6G 17G 18% /
none 256M 4.0K 256M 1% /dev
4) Yes

Well it was just a test container and it's destroyed now. In the meantime I changed the hardware node from Centos 6.0 to CL6.1, hopefully that'll take care of any issues.

Report message to a moderator

Re: Unable to stop container [message #43619 is a reply to message #43574]

Sun, 02 October 2011 13:29

Ales
Messages: 330
Registered: May 2009

Senior Member

A typo, yes. Should have said "/sbin/iptables-save and /sbin/iptables-restore". From what you have written, everything seemed normal on your previous system... Odd.

Report message to a moderator