[Lxc-users] start order

Michael H. Warfield mhw at WittsEnd.com
Sat Dec 8 19:59:56 UTC 2012


On Sat, 2012-12-08 at 19:51 +0100, Daniel Baumann wrote:
> On 12/08/2012 06:24 PM, Michael H. Warfield wrote:
> > It's possible to drop a full configuration file there into /etc/lxc/auto and
> > not have it exist in /var/lib/lxc and then "-f <config file>" would work
> > while "-n  $c -d" would fail.  That would be my guess there.

> that was the intention when i wrote it in debian (ubuntu at some point 
> took it over then).

> > I've been imitating the way OpenVZ does this (on Fedora) by adding some
> > parameters in the lxc configuration files like this:

> not convinced that a start order does actually matter (because your 
> stuff within containers should a) be graceful wrt/ to existence of other 
> services by other containers and b) be eventful), however, better and 
> simpler imho is to support using numbered prefixes.

I can see his desire for controlling start order.  In my "real world"
case, where I have to reboot one of my major hosts, I have certain
important containers I need up as soon as possible.  One is a null
router which is the /dev/null sink for my net-telescope (security
related stuff similar to a honeypot).  That's the first one that needs
to be up.  That one is followed by some authoritative name servers and
recursive name servers.  After that, there are processes which like to
chew on MySQL databases as they slowly come up.  The containers are
independent of one another but they're demand on the host resources can
have a significant real-time impact.  Not vital.  Everything comes up
anyways, but I can see his point.

> /etc/lxc/auto/0001-foo
> /etc/lxc/auto/0002-bar
> [...]

I like having the containers named for the host names (the old OpenVZ
paradigm really sucked) so listing give me better readability and
understandability.  I thought of the prefix idea too but discounted it
because it would then force you into a naming convention for the
containers themselves.  That's more of a hack than a solution in that
case where you have additional side effects like that.

> -- 
> Address:        Daniel Baumann, Donnerbuehlweg 3, CH-3012 Bern
> Email:          daniel.baumann at progress-technologies.net
> Internet:       http://people.progress-technologies.net/~daniel.baumann/

Regards,
Mike
-- 
Michael H. Warfield (AI4NB) | (770) 985-6132 |  mhw at WittsEnd.com
   /\/\|=mhw=|\/\/          | (678) 463-0932 |  http://www.wittsend.com/mhw/
   NIC whois: MHW9          | An optimist believes we live in the best of all
 PGP Key: 0x674627FF        | possible worlds.  A pessimist is sure of it!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 482 bytes
Desc: This is a digitally signed message part
URL: <http://lists.linuxcontainers.org/pipermail/lxc-users/attachments/20121208/329ec1e8/attachment.pgp>


More information about the lxc-users mailing list