OpenVZ Forum


Home » Mailing lists » Users » VE fails to stop
Re: VE fails to stop [message #25084 is a reply to message #25069] Fri, 14 December 2007 09:35 Go to previous messageGo to previous message
kir is currently offline  kir
Messages: 1645
Registered: August 2005
Location: Moscow, Russia
Senior Member

Cliff Wells wrote:
> On Thu, 2007-12-13 at 19:02 +0300, Kir Kolyshkin wrote:
>   
>> OK, here's what I found concerning ubuntu-7.10 VE vs. OpenVZ.
>>     
>
> Seems to work fine.  Minor nit would be the unknown group 11.  I did a
> 'find -gid 11' and it turned up a bunch of /dev/fd* stuff with that GID,
> but chowning them to root didn't get rid of the message.
>   

This message ("id: cannot find name for group ID 11") comes from id 
binary which is executed from /etc/profile during user login (or vzctl 
enter). Since root belongs to group 11, "id" tries to get the name of 
the group.

The line about group with ID of 11 is absent from /etc/group. I'm not 
sure why. Since you found out some /dev/fd* files belong to the group, I 
guess group name should be something like 'floppy'. I guess that adding 
something like
floppy:x:11:
to /etc/group should fix the issue.

I will file a bug to ubuntu about that, but generally it should be 
harmless as it is.

> Networking now seems to work fine.
>
> I was a bit curious about the results of 'mount':
>
> vps1 ~ # vzctl enter 108
> entered into VE 108
> id: cannot find name for group ID 11
> root@vz108:/# mount
> simfs on / type simfs (rw,noatime)
> proc on /proc type proc (rw,nosuid,nodev,noexec)
> sysfs on /sys type sysfs (rw,nosuid,nodev,noexec)
> tmpfs on /var/run type tmpfs (rw,nosuid,nodev,noexec)
> tmpfs on /var/lock type tmpfs (rw,nosuid,nodev,noexec)
> tmpfs on /dev/shm type tmpfs (rw)
> devpts on /dev/pts type devpts (rw)
> tmpfs on /var/run type tmpfs (rw,nosuid,nodev,noexec)
> tmpfs on /var/lock type tmpfs (rw,nosuid,nodev,noexec)
> root@vz108:/# 
>
> It looks as if perhaps the tmpfs's are being mounted twice now?
>   
Yep, first time they do it in /etc/init.d/mountkernfs.sh, then in 
/etc/init.d/mountall.sh (calling pre_mountall function from sourced 
/lib/init/mount-functions.sh) they bind-mount both /var/run and 
/var/lock to under /dev/shm. That is why we see what we see.

I am not touching that stuff now because I already learned they assume a 
lot of things (like /var/run is clean after reboot) and one can open a 
can of worms breaking their assumptions.

Still I don't understand why they mount twice and I will file a bug 
about it.
> Overall this is much better than it was before.
>   

Hope so :)
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: Virtual Ethernet device
Next Topic: Cloning and permissions
Goto Forum:
  


Current Time: Tue Jul 01 08:50:48 GMT 2025

Total time taken to generate the page: 0.05940 seconds