[Lxc-users] Very odd lxc behavior

[David Parks]

Vanilla 12.10 Ubuntu server:

 

root at atlas01:~# sudo lxc-version

lxc version: 0.8.0-rc1

 

I may very well have accidentally tried to start the same container twice, I
think I fat-fingered something just before encountering the issue, and I've
never had it happen on any of our other boxes. Anyway, it's of no real
consequence to me now as I cleared it easily enough, but I thought it worth
mentioning.

 

Dave

 

 

-----Original Message-----
From: Serge Hallyn [mailto:serge.hallyn at ubuntu.com] 
Sent: Friday, April 12, 2013 8:58 PM
To: David Parks
Cc: lxc-users at lists.sourceforge.net
Subject: Re: [Lxc-users] Very odd lxc behavior

 

Quoting David Parks ( <mailto:davidparks21 at yahoo.com>
davidparks21 at yahoo.com):

> Today I saw this.

> 

> I had started half the containers when I noticed the problem (not all 

> of them as the output below suggests).

> 

> I stopped them all (which generated some errors, seen at the end).

> 

> After stopping them all everything went back to normal and I could 

> start them all successfully without "strangeness".

> 

>  

> 

>  

> 

> root at atlas01:/var/lib/lxc# lxc-list

> 

> RUNNING

> 

> ls: cannot access testdb1: No such file or directory

 

Is this by chance an older version of lxc?  It sounds like there were later
attempts to start testdb1, and so

/sys/fs/cgroup/freezer/lxc/testdb1 got renamed to deadXXXXX, then the later
attempt failed, so

/sys/fs/cgroup/freezer/lxc/testdb1 ended up not existing.

 

Note that that behavior no longer exists in very recent lxc.

Rather, the second attempt would use

/sys/fs/cgroup/freezer/lxc/testdb1-2, and lxc-list would ask the container
monitor for the right cgroup path, preventing this -ENOENT.  At least that's
the theory :)

 

-serge

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxcontainers.org/pipermail/lxc-users/attachments/20130413/b0e279a0/attachment.html>