<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Thanks a lot Stig... your analysis in that discussion goes way
beyond mind. Did you ever sort it out?<br>
<br>
In the meantime I found an unlikely setting that solved my problem:
the session_linger. I got it from here [1] and thought it wouldn't
hurt and it nearly killed me. My initial tests were conducted with a
value of 150, I had to lower it to 20 to get my test through.<br>
<br>
Thanks + best regards<br>
<br>
[1]
<a class="moz-txt-link-freetext" href="https://kristianlyng.wordpress.com/2009/10/19/high-end-varnish-tuning/">https://kristianlyng.wordpress.com/2009/10/19/high-end-varnish-tuning/</a><br>
<br>
On 05.01.2011 23:56, Stig Bakken wrote:
<blockquote
cite="mid:AANLkTincEpY0fOWmLR-FmdgLWD3dPYPUQRPjAAzPEFTv@mail.gmail.com"
type="cite">This thread: <a moz-do-not-send="true"
href="http://www.varnish-cache.org/lists/pipermail/varnish-misc/2010-December/005258.html">http://www.varnish-cache.org/lists/pipermail/varnish-misc/2010-December/005258.html</a>
<div><br>
</div>
<div> - Stig<br>
<br>
<div class="gmail_quote">On Wed, Jan 5, 2011 at 11:44 PM, George
Georgovassilis <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:g.georgovassilis@gmail.com">g.georgovassilis@gmail.com</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt
0.8ex; border-left: 1px solid rgb(204, 204, 204);
padding-left: 1ex;">
<div bgcolor="#ffffff" text="#000000"> Hello Stig,<br>
<br>
Thanks for the insight. I'm still on the logs, though not
sure where to start - it's not like that there are any
errors in it so I'm not really sure what to look for. Do
you have a pointer to that discussion you are referring
to?
<div>
<div class="h5"><br>
<br>
On 05.01.2011 23:41, Stig Bakken wrote:
<blockquote type="cite">This seems similar to what
I've been seeing, described in an earlier thread
from before christmas. In my case it was not during
benchmarking, but when serving production load of
around 300 req/s per server. Modern tcpip stacks on
modern hardware should handle this without blinking.
<div> <br>
</div>
<div>Did you have the chance to capture the problem
with varnishlog so you can replay/analyze it?</div>
<div><br>
</div>
<div> - Stig<br>
<br>
<div class="gmail_quote">On Wed, Jan 5, 2011 at
11:18 PM, George Georgovassilis <span dir="ltr"><<a
moz-do-not-send="true"
href="mailto:g.georgovassilis@gmail.com"
target="_blank">g.georgovassilis@gmail.com</a>></span>
wrote:<br>
<blockquote class="gmail_quote" style="margin:
0pt 0pt 0pt 0.8ex; border-left: 1px solid
rgb(204, 204, 204); padding-left: 1ex;">I
removed the varnish instance so that the load
generator is directly hitting Tomcat.
Naturally, the request rate drops to 70
requests/sec with a CPU load of 100%...
however connections don't drop anymore, no
timeouts occur and the application remains
pretty responsive. To recap, these are the
possible scenarios:<br>
<br>
1. The networking layer is overtaxed with the
original 300 reqs/sec. I don't believe that,
because the load generator doesn't record any
dropped connections while a simple browser
can't connect.<br>
<br>
2. Tomcat is overtaxed. That also seems not
plausible, since it is not servicing any
requests under the load test - all is done by
varnish. Even if, as I said when removing
varnish from in between, it serves the
requests just fine.<br>
<br>
3. Varnish is overtaxed. Somehow that also
doesn't make sense, since it is servicing the
load generator just fine... but will refuse to
serve browser requests.<br>
<br>
4. Varnish, when under load, is picky about
what connections to serve.<br>
<br>
I'm stuck :-)
<div>
<div><br>
<br>
On 05.01.2011 17:59, Bob Camp wrote:<br>
<blockquote class="gmail_quote"
style="margin: 0pt 0pt 0pt 0.8ex;
border-left: 1px solid rgb(204, 204,
204); padding-left: 1ex;"> Hi<br>
<br>
Running simple load tests both on Apache
directly, and on Varnish - both<br>
seem to experience "long delays" on a
small percentage of the requests. The<br>
problem does not appear to happen with
low loads. It does come up as CPU<br>
usage becomes an issue. It also is hard
to make happen with a single stream<br>
of requests. It seems to come up much
quicker with many requests done in<br>
parallel.<br>
<br>
I've always *assumed* that the poor
little TCP/IP hamster simply ran out of<br>
breath and started dropping connections.<br>
<br>
Bob<br>
<br>
-----Original Message-----<br>
From: <a moz-do-not-send="true"
href="mailto:varnish-misc-bounces@varnish-cache.org"
target="_blank">varnish-misc-bounces@varnish-cache.org</a><br>
[mailto:<a moz-do-not-send="true"
href="mailto:varnish-misc-bounces@varnish-cache.org"
target="_blank">varnish-misc-bounces@varnish-cache.org</a>]
On Behalf Of George<br>
Georgovassilis<br>
Sent: Wednesday, January 05, 2011 11:18
AM<br>
To: <a moz-do-not-send="true"
href="mailto:varnish-misc@projects.linpro.no"
target="_blank">varnish-misc@projects.linpro.no</a><br>
Subject: Re: Connections dropped under
load<br>
<br>
Hello Cosimo,<br>
<br>
Thank you for the quick reply. After
your hint I had the tests run again<br>
but couldn't detect that pattern. What
susprised me though after looking<br>
through the logs is that almost all
requests by the load generator<br>
complete in a timely manner (< 1
sec), but all requests generated by a<br>
real browser (IE, FF, Opera) will be
served much later or even run into<br>
a timeout.<br>
<br>
On 05.01.2011 16:30, Cosimo Streppone
wrote:<br>
<blockquote class="gmail_quote"
style="margin: 0pt 0pt 0pt 0.8ex;
border-left: 1px solid rgb(204, 204,
204); padding-left: 1ex;"> On Wed, 05
Jan 2011 16:20:31 +0100, George
Georgovassilis<br>
<<a moz-do-not-send="true"
href="mailto:g.georgovassilis@gmail.com"
target="_blank">g.georgovassilis@gmail.com</a>>
wrote:<br>
<br>
<blockquote class="gmail_quote"
style="margin: 0pt 0pt 0pt 0.8ex;
border-left: 1px solid rgb(204, 204,
204); padding-left: 1ex;"> I'm
having trouble with dropped
connections under a loadtest.<br>
<br>
The problem: As a measure for
response, I am requesting an image
from<br>
the webapp running in Tomcat while
the loadtest is underway. However<br>
that either times out or is
delivered after several seconds.
Varnishlog<br>
will often either not show the
request (RxURL) at all, or show it<br>
several seconds after the browser
dispatched it.<br>
</blockquote>
Hi George,<br>
<br>
if you measure the time you mention as
"several seconds"<br>
and it's either 3 or 9 seconds, I
think what you're seeing<br>
is a client-side TCP retransmit
timeout.<br>
<br>
I experienced that, both under load
testing,<br>
and in real production setups.<br>
<br>
</blockquote>
<br>
_______________________________________________<br>
varnish-misc mailing list<br>
<a moz-do-not-send="true"
href="mailto:varnish-misc@varnish-cache.org"
target="_blank">varnish-misc@varnish-cache.org</a><br>
<a moz-do-not-send="true"
href="http://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc"
target="_blank">http://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc</a><br>
<br>
</blockquote>
<br>
<br>
_______________________________________________<br>
varnish-misc mailing list<br>
<a moz-do-not-send="true"
href="mailto:varnish-misc@varnish-cache.org"
target="_blank">varnish-misc@varnish-cache.org</a><br>
<a moz-do-not-send="true"
href="http://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc"
target="_blank">http://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc</a><br>
</div>
</div>
</blockquote>
</div>
<br>
<br clear="all">
<br>
-- <br>
Stig Bakken<br>
CTO, Zedge.net - free your phone!<br>
</div>
</blockquote>
<br>
</div>
</div>
</div>
<br>
_______________________________________________<br>
varnish-misc mailing list<br>
<a moz-do-not-send="true"
href="mailto:varnish-misc@varnish-cache.org">varnish-misc@varnish-cache.org</a><br>
<a moz-do-not-send="true"
href="http://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc"
target="_blank">http://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc</a><br>
</blockquote>
</div>
<br>
<br clear="all">
<br>
-- <br>
Stig Bakken<br>
CTO, Zedge.net - free your phone!<br>
</div>
</blockquote>
<br>
</body>
</html>