[ec2ubuntu] seeing high load on brand new lucid instance (postgres issue?)

26 views
Skip to first unread message

James Turk

unread,
Apr 30, 2010, 1:05:29 PM4/30/10
to ec2ubuntu
I'm fairly new to EC2 so forgive me if I'm missing something obvious.

If I start up a fresh server based on the new 10.04 64-bit image
(ami-4b4ba522) and instal postgres-8.4 I quickly see the load average
go up to somewhere consistently between 0.7 and 1.2.

I've also tried a non-EBS version of the same image (ami-714ba518) and
as soon as i install postgres the load goes up and doesn't come down.
Just to be clear, I haven't even created a database on these servers,
and running top shows no CPU utilization.

After stopping postgres load drops rather quickly (down to 0.05 or so
like I'd expect). I've spun up a few instances to test on and each
one has shown the same behavior.

I'm reporting it here because I don't see similar behavior on my
desktop machine that is also running Lucid with postgres installed so
I'm inclined to believe it may be something unique to the AMIs or EC2.

Do others running Lucid on EC2 see similar issues?

-James

--
You received this message because you are subscribed to the "ec2ubuntu" Google Group:
http://groups.google.com/group/ec2ubuntu
To unsubscribe, send email to ec2ubuntu-...@googlegroups.com

Rod

unread,
May 3, 2010, 9:27:34 PM5/3/10
to ec2ubuntu

Hi James,
Didn't see your message before I posted mine James, maybe it hadn't
come through moderation. I'm having very similar issues on 32-bit but
not involving postgres, but I do have MySQL, not sure if that's
related to my problems but I might have a play with a new instance.
See http://groups.google.com/group/ec2ubuntu/browse_thread/thread/4be26e81b7c597bc
-- Rod

James Turk

unread,
May 3, 2010, 9:37:46 PM5/3/10
to ec2u...@googlegroups.com
Rod,

Just saw your post too, seeing nearly identical numbers.

I've tried out a number of instances and it happens for me each time,
what I'd be curious is if there is anyone that *isn't* seeing these
weird
load numbers.

I also had a conversation on IRC with someone in #ubuntu-server and
after showing similar numbers it was suggested this was likely a weird
artifact of the VM, it is still strange and seems like a regression
since 9.10.

Pete Jalajas

unread,
May 3, 2010, 11:09:53 PM5/3/10
to ec2ubuntu
Hi all,
Not at all sure if it's related, but, for several weeks now (launched
2010-03-10 11:09 EST), I've been using an Ubuntu Lucid 10.04 AWS EC2
ami-d5cc23bc instance, apt-get dist-upgrade'ing several times a week
(if not several times a day), as a remote desktop with x2go, as a test
replacement for my aging basement desktop server, and it has been
working amazingly well, that is, until a few days ago. Recently it
has become unusably slow and the x2go session just goes white if I
leave it for many minutes, after which I have to terminate the x2go
session. It seemed to be working _much_ better today, in fact, I did
several hours of real billable work on it today, but I've just
returned from my Monday night CTU addiction, and now the x2go white
screen ("not responding") greets me.
Sorry if this is irrelevant,
Pete

Scott Moser

unread,
May 4, 2010, 10:14:22 AM5/4/10
to ec2u...@googlegroups.com
On Mon, 3 May 2010, James Turk wrote:

> Rod,
>
> Just saw your post too, seeing nearly identical numbers.
>
> I've tried out a number of instances and it happens for me each time,
> what I'd be curious is if there is anyone that *isn't* seeing these
> weird
> load numbers.
>
> I also had a conversation on IRC with someone in #ubuntu-server and
> after showing similar numbers it was suggested this was likely a weird
> artifact of the VM, it is still strange and seems like a regression
> since 9.10.

James,
Could you please open a launchpad bug for this ? Ideally do so running
'ubuntu-bug linux-ec2' on the instance. I apologize for what might seem like
busywork, but before we can actually fix it we'll need a bug.
I've pointed John Johansen at this thread [1] and the other [2].

Thanks for finding this issue and helping us resolve it.

[1] http://groups.google.com/group/ec2ubuntu/browse_thread/thread/a7e9bc45cf923f8c
[2] http://groups.google.com/group/ec2ubuntu/browse_thread/thread/4be26e81b7c597bc

James Turk

unread,
May 4, 2010, 10:33:15 AM5/4/10
to ec2u...@googlegroups.com
I took a stab at a bug report
https://bugs.launchpad.net/ubuntu/+source/linux-meta-ec2/+bug/575193,
let me know if there's anything else I should include.

Rod- I included a reference to your issues but you have some good
numbers in yours that might be useful to attach to the report.

-James

Rod

unread,
May 14, 2010, 12:20:19 AM5/14/10
to ec2ubuntu

Scott, are you able to tell us about progress at looking into this
issue? Some feedback would be great and please let us know if there is
any more information we can provide. James & I are subscribed to the
Launchpad bug report.
Much appreciated,
-- Rod

MS

unread,
May 19, 2010, 1:28:06 PM5/19/10
to ec2ubuntu
I am also seeing this issue. My system seems pretty responsive even
with this load.

cat /proc/loadavg
5.43 5.38 5.37 1/91 5330

I am wondering if this is just a reporting issue. I don't want to go
back to Karmic :-/

Marius

Pete Jalajas

unread,
May 19, 2010, 2:09:53 PM5/19/10
to ec2u...@googlegroups.com
FWIW, I'm using a Ubuntu 10.04 Lucid AWS EC2 small instance as a
remote desktop and it works great. It slowed down for several days,
but I tracked it down by running
sudo top -c
and then sorting by RAM (Shift-m) and seeing polkitd hogging RAM. I
then tracked that down to pulseaudio flailing, and so I deleted my
~/.pulseaudio per
https://bugs.launchpad.net/bugs/570015
and it has been very snappy ever since.
Google Chrome is my current RAM hog, but doesn't seem to be getting in
the way.
BTW:
cat /proc/loadavg
0.02 0.21 0.26 1/244 16234

On Wed, May 19, 2010 at 1:28 PM, MS <lov...@gmail.com> wrote:
> I am also seeing this issue. My system seems pretty responsive even
> with this load.
>
> cat /proc/loadavg
> 5.43 5.38 5.37 1/91 5330
>
> I am wondering if this is just a reporting issue. I don't want to go
> back to Karmic :-/
>
> Marius
>

Reply all
Reply to author
Forward
0 new messages