[Systems] Load on treehouse

[Sascha Silbe]

Excerpts from Bernie Innocenti's message of Thu Sep 02 13:26:40 +0200 2010:

[IPv6 to treehouse]
> Works perfectly for me. I tested ping, http and ssh from treehouse and
> bender.
OK, then it's the SixXS tunnel. Looks like I need to change the PoP;
it's been a bit flaky in general for the past few months. The tunnel
on my laptop (SixXS too, but different PoP) is rock-stable OTOH (or at
least as stable as it can be, given the underlying IPv4 network).

> I've now installed atop, so we can go back in the past and see what's
> wrong. Does sks need much memory? Could it make the machine swap?
I've had it running before on a machine with 384MB RAM (physical), so
it shouldn't need several GBs.
lightwave doesn't even have swap configured. I've seen this cause
problems (high CPU load) on several systems over many years, so my
first suggestion would be to enable swap, even if just a minimal amount
(say 32MB).

> Anyway, I was planning to upgrade lightwave to lucid (sounds like a pun,
> eh?). Shouldn't be hard, as it runs just named and sks. We'll re-measure
> performance after the upgrade.
BIND only has 2 open bugs in Ubuntu [1]. That's suspiciously low, but no
reason not to upgrade. SKS is backported from Lucid already, so not an
obstacle either.

> Of course. And anyway, ionice cannot do miracles even in non-virtualized
> case due to ordering constraints of journaling filesystems. Moreover,
> seek time dominates and the Linux block layer only cares about I/O
> bandwidth.
I'd expect the statistics calculation to be entirely read-only which
should help quite a bit.

> As of 2010, fair I/O scheduling at the OS level is still a largely
> unsolved problem (researchers thought they had solved it in publications
> of the 70's, but they forgot to tell us how to implement it in a real
> kernel with a real filesystem and a real workload :-)
:-/

> > 2. running high-priority (LDAP) and low-priority (SKS) servers inside
> >    the same VM (so we can't just ionice the entire VM).
> the LDAP server is still not running on lightwave.
Ah, good to know.

> Maybe I'm paranoid, but even after 175 days of uptime I still don't
> trust treehouse's stability enough.
I thought the plan was to have a backup LDAP server on sunjammer? So
why does lightwave/treehouse need to be ultra-reliable?
(The answer might very well be "because LDAP sucks" - I have no idea).

> There are a number of pending infrastracture changes that I'm planning
> to do after I'll be settled in Boston because they are too risky from
> remote.
Interesting. What in particular?
BTW, you still have my +1 on the fat server if you think it's useful.
I'm not quite convinced about it (especially given problems like those
mentioned above), but will stand behind your decision either way.

> > So maybe we should move SKS into a new VM or at least a different,
> > equally low-priority one (that we will ionice).
> There are no priorities for VMs. Neither I/O, nor CPU.
I thought KVM is more or less a regular user process, with (almost)
all applicable restrictions? Do you have a pointer to some documentation
explaining the issue?

Sascha

[1] https://bugs.launchpad.net/ubuntu/+source/bind/+bugs
-- 
http://sascha.silbe.org/
http://www.infra-silbe.de/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 490 bytes
Desc: not available
Url : http://lists.sugarlabs.org/private/systems/attachments/20100902/5d772a0f/attachment.pgp