I have two mostly idle Xen VMs, A and B. Nagios is alerting of 10
second timeouts when GETting a static file from A (client) to a vhost
on B, but this affects only some virtual servers. The files are static
HTML and we're not requesting any referenced files.

The issue only appears to be between these two hosts - requesting the
same file from my ADSL connection shows no such delay.

The delay is quite reliable; HTTP GETs for a tiny file on a delayed
host come in around the 10.008s mark, while GETs on a non-delayed host
show consistent times which are 10,000 ms less.

Apache on B is Apache2-mpm-prefork 2.2.4-3ubuntu0.1. SSI is turned
off, as are DNS lookups for logging. I'm not using GeoIP on this host.

Apache configs for a site which does delay, and another site which
reliably doesn't delay, are identical in all respects except for
hostname and directories.

Has anyone seen something similar? My first thought was DNS, but I'm
suspecting it's actually related to the network between the Xen hosts.

There's no measurable delay in looking up the requesting hostname from
the server (I've added it to /etc/hosts now, but that made no change).

What might introduce a 10,000ms delay to HTTP?

--~--~---------~--~----~------------~-------~--~----~
NZ PHP Users Group: http://groups.google.com/group/nzphpug
To post, send email to [email protected]
To unsubscribe, send email to
[email protected]
-~----------~----~----~----~------~----~------~--~---

Reply via email to