[lxc-users] lxc insttability

[Király, István]

Hello list, ..

I'm wondering if anyone experienced this ...

Randomly after some operation, eg, starting / stopping / adding containers,
or some software installation one-two containers start being errorrous. ssh
is running in the container, but connections from the host get refused, nfs
that worked stops to work, errno 111 (Connection refused), webservers go
down, .. Ping responds usually fast, and lxc-attach works, but then ...
slowly the number of faulty containers grow, until all containers get
unresponsive, and at the end even the host can get unresponsive. After the
host is rebooted everything starts up, and works, and things operate again
normal for months, .. with 120 containers or even more.

I'm wondering where the bug might be, and how to find it. I use the default
libvirt network, maybe the bug is over there, but since i didnt find a way
to reproduce this, Im kind of tapping in the dark. On my development server
it happens often, where I change configurations frequently ... but needless
to say, sometimes that can hhappen on production-like systems .. :/
I use the packaged fedora lxc now. I know that the dev is more taking place
on ubuntu, but I would like to stick to fedora.

So anyone saw anything similar?

Greetings, ...



-- 
 Király István
+36 209 753 758
LaKing at D250.hu
<http://d250.hu/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linuxcontainers.org/pipermail/lxc-users/attachments/20160219/a936eb0c/attachment.html>