Thank you for the tip regarding strace. From a first glance it looks like server 2 loses time between "Writing spool header file" and "LOG: MAIN unfrozen by root". Especially some fstat(), fsync(), close() call lines have higher times displayed when using strace with the -r param.
E.g. one close() shows 0.000644 on server 1 and the same close() on server 2 shows 0.032874.
As for the file systems: none of them is on LVM. Server 1 uses ext3, server 2 uses ext4.
So my uneducated guess would be that ext4 is to blame. Is there a way to temporarily make ext4 behave like ext3 and re-run the test?
Thanks,
Volker
-----Original Message-----
From: Todd Lyons [
mailto:tlyons@ivenue.com]
Sent: Wednesday, August 22, 2012 3:17 PM
To: Volker Schmelich
Cc: exim-users@???
Subject: Re: [exim] Exim message thaw performance
On Wed, Aug 22, 2012 at 10:34 AM, Volker Schmelich <volker@???> wrote:
> Unfortunately, we are talking about different hardware, different OS,
> different Exim versions.
One test you can do if you have the hardware available is to install the new versions on identical hardware and make sure that it's not some I/O subsystem which causes it. However, I suspect it's not the hardware.
> Could it be caused by changes in between Exim versions - internal
> processing changes?
> Could it be caused by 32 vs. 64 bit or different kernels?
> Any suggestions where to start or which tools could be used to gain
> further insight?
The only real difference I notice is a different version of Berkeley DB.
Are the filesystems the same? Is one using ext3 and the other using ext4? Is one on LVM and the other is not on LVM?
"strace -f -tt -T" for each queue run. Look for differences in time spent loading shared libs. Look for differences in time spent in each function call. Hopefully something in that output will give you an idea of where the difference in time is spent.
...Todd
--
The total budget at all receivers for solving senders' problems is $0.
If you want them to accept your mail and manage it the way you want, send it the way the spec says to. --John Levine
