Hi Lukas,
On 05.04.2013 12:00, Lukas Tribus wrote:
Whats is the percentage of requests failing this way?
I'm not sure. But I think it's less than 1%. We do a couple of 100's
request per second and about every second I see one failed request.
Do you know if this is an issue introduced by a certain haproxy build,
and thus
was working previously, or did you only recently enable splice-auto?
Are you able to reproduce this in dev17 or in stable 1.4.23 (but you
probably
rely on 1.5 features)?
I cannot try 1.4 because I indeed rely on 1.5 features. But I did try
dev18 and dev17-ss-20130125. Both give the same problems. I cannot go
any further back because I had some issues with versions before 20130125
if I recall correctly. I'm not sure what is was anymore :-)
Can you remove splice-auto, and check whether splice-request or
splice-response
or both are affected?
Using splice-request and splice-response I get the same issue.
Using splice-request gives no problems.
Using splice-response I get the issue again.
Do you see this in a lab setup as well or do you need to troubleshoot
this
with production services?
I do not have a big lab setup in which I can reproduce this.
Are you able to tcpdump an affected session (both front and backend
traffic)?
It is possible to do that, but only if really necessary. And I probably
only want to share that with direct HAProxy developers.
I use kernel 3.2.40 with grsec patch
Any kernel messages in dmesg?
Nope, not anything out of the ordinary.
Do you have the possibility to install a stable but recent vanilla
kernel
from kernel.org (I suppose 3.8.5 would be a good choice)? This may as
well
be a kernel issue.
Vanilla 3.2.X would be possible, anything else is a bit more
problematic. Not impossible, but I only want to do that if everything
else fails.
Greets,
Sander
