Can't Stop/Start/Remove a VPS

[SkylarM]

So I have a VPS that I need to remove/delete from the node. I go into my control panel (solus) and try to do it the easy way, and it deletes it from Solus, but OpenVZ shows it as online and kicking away still. So I SSH in and try to stop the VPS manually and see what's going up.
 

 

This is what I get:

 

vzctl stop 150
Stopping container ...
Child 149193 exited with status 7
Killing container ...
Child 149260 exited with status 7
Unable to stop container

 

 
Mind you the Child ID changes each time I run the command, so it's not a simple kill the PID.
So okay I'll try to enter it and shut it down manually.

 

vzctl enter 150
enter into CT 150 failed

 

 

I've looked everywhere I can think of and I just cannot seem to find a solution to this. I can't enter it, I can't change the root password, I can't stop/restart/delete it. Basically can't do anything other than stare at it. This literally has me stumped for the first time in a very long time. Any suggestions appreciated!

 

[jarland]

Try and remove the IP from the container, consider the container dead until you have to reboot. Maybe someone else here has some insight that I'd be equally grateful for, but I've not had much luck in these cases. Tracking down the processes and killing them hasn't been much help to me the last 3 times I've had it happen.

 

[SkylarM]

Try and remove the IP from the container, consider the container dead until you have to reboot. Maybe someone else here has some insight that I'd be equally grateful for, but I've not had much luck in these cases. Tracking down the processes and killing them hasn't been much help to me the last 3 times I've had it happen.

The problem is the container is sitting at a solid 1.0 CPU load no matter what I do, which is really what I want to try and fix.

So what you just set the ve config to disabled and when you rebooted removed it then?

 

[mikho]

From what I read on the internet, using a famous search engine, the suggestions are to restart the node.

 

[SkylarM]

From what I read on the internet, using a famous search engine, the suggestions are to restart the node.

I've obviously seen similar results, was hoping for another potential workaround that I wasn't seeing online.

 

[mikho]

Just saying that every time I read about this, the only solution was to reboot the node.


Couldn't find any other solution, not even working for people who killed almost every process inside the container.


So it looks like the only solution. Perhaps you have better luck!

 

[SkylarM]

Ah well. Have to take the server down Saturday to move it to the new cab anyways so guess I'll deal with it then

 

[SeriesN]

Try restarting ovz instead of the whole node.

 

[HalfEatenPie]

Not too sure, but have you tried any of these?

http://forum.openvz.org/index.php?t=msg&goto=27976&

http://serverfault.com/questions/318711/stop-an-unstoppable-openvz-container

http://forum.proxmox.com/threads/12587-Problems-with-shutting-down-container

Instead of restarting the node, if you know the vzctl container ID then have you checked those ways?

Yeah I know that you don't want to restart the entire node or anything.  Trying to find different ways to fix it.  

 

[KuJoe]

Kill the init process?

 

[acd]

Agree with KuJoe. vzps -E 150... if there's anything in there, kill it with fire (vzctl stop 150 --fast). If there isn't, file a bug with ovz.

 

[Magiobiwan]

If you kill the VPS's init process (kill -9 thepid) the VPS will stop. You just have to make sure you kill the right init, otherwise some poor fellow's VPS will crash mysteriously. 

 

[Zen]

Damn. I'm too late. But yeah what everyone said, just kill the init process.

 

[ICPH]

Kill the init process?

How to kill it while VM returns "enter into CT 190 failed" and "Unable to stop container" ?

 

[KuJoe]

How to kill it while VM returns "enter into CT 190 failed" and "Unable to stop container" ?

You can kill it from the host node, no need to login to the VPS to kill VPS processes with OpenVZ.

 

[ICPH]

Thx, i tried:

vzctl exec 190 kill pidhere

vzctl exec 190 kill -9 pidhere

vzctl exec 190 kill *

but when i do:

vzctl restart 190

but it appears it always report another pid ...

(Child PIDHERE exited with status 7)

any idea please?