Quoting Andreas Hasenack ([email protected]):
> On Tue, May 27, 2014 at 2:22 PM, Serge Hallyn <[email protected]>wrote:
>
> > Quoting Andreas Hasenack ([email protected]):
> > > Hi,
> > >
> > > for an experiment I'm trying to deploy 100 containers on a machine. I
> > used
> > > the "ubuntu" charm and basically did a for loop with juju add-unit ubuntu
> > > --to lxc:1
> > >
> > > I had to adjust some ulimit limits, but so far, so good.
> > >
> > > It's taking more time than I expected to deploy this, but what really
> > > caught my attention is how long lxc-ls --fancy takes to complete:
> > >
> > > # time lxc-ls --fancy | grep juju-machine-1|wc -l
> > > 91
> > >
> > > real 1m36.950s
> >
> > the --fancy does add a lot of work, but this seems slower than it ought
> > to be. Are the containers running while you are doing this? Is it
> > faster if they happen to all be stopped or all started?
> >
>
>
> They were running, but doing nothing. It was just a deployment of the
> "ubuntu" charm. Didn't try with them all stopped, and right now I don't
> have that system up anymore.
Hi,
I just tried this here now. I used 100 overlayfs clones, started them
all, and did
time sudo lxc-ls -P /mnt/lxc -f
real 0m3.448s
user 0m0.397s
sys 0m0.594s
Now maybe this has to do with the types of containers you are using. Can
you try the following and see if you get the results I do?
sudo lxc-create -t download -n c-1 -- -d ubuntu -r trusty -a amd64
for i in `seq 1 100`; do
sudo lxc-clone -s -o c-1 -n clone-$i
done
for i in `seq 1 100`; do
sudo lxc-start -n clone-$i -d
done
time sudo lxc-ls -f
-serge
--
ubuntu-server mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-server
More info: https://wiki.ubuntu.com/ServerTeam
