Mailing List Archive

[Bug 517] Load average variable stuck during long connections
------- You are receiving this mail because: -------
You are the QA contact for the bug, or are watching the QA contact.

http://www.exim.org/bugzilla/show_bug.cgi?id=517





------- Comment #1 from ph10@hermes.cam.ac.uk 2007-06-25 11:07 -------
On Fri, 22 Jun 2007, moseleymark@gmail.com wrote:

> The spam filter box keeps connections open for a long time to save on
> session start up (it'll keep it open for a tunable amount of emails or
> till Exim times it out).

This is what is causing your problem. Exim makes the assumption that
connections do not last particularly long, and that they are from a
random collection of hosts sending it messages.

> I'm seeing something odd: Exim is queueing incoming messages from the spam
> filter box due to high load even when the load is low. All I can guess is that
> the initial load average gets reused over and over, despite the fact that
> /proc/loadavg is rechecked for each message. Or perhaps once it hit the
> 'queue_only_load' threshold, some flag got set that hasn't gotten unset.

The latter. This is the comment in the code:

Note that, once set, local_queue_only remains set for any subsequent
messages on the same SMTP connection. This is a deliberate choice; even
though the load average may fall, it doesn't seem right to deliver later
messages on the same call when not delivering earlier ones.

> connection in question has been connected for about 6 hours!

...way outside the sort of connections I have ever considered...


So what is to be done? I suppose the only answer is to add an option to
change the behaviour so that it recalculates the load average for every
message.

Philip

--
Configure bugmail: http://www.exim.org/bugzilla/userprefs.cgi?tab=email

--
## List details at http://www.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##
[Bug 517] Load average variable stuck during long connections [ In reply to ]
------- You are receiving this mail because: -------
You are the QA contact for the bug, or are watching the QA contact.

http://www.exim.org/bugzilla/show_bug.cgi?id=517


ph10@hermes.cam.ac.uk changed:

What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED




------- Comment #2 from ph10@hermes.cam.ac.uk 2007-06-27 12:05 -------
1. As a grumpy old man, I should also point out that if you had read the
documentation for queue_only_load, you would have found this sentence: "If this
happens during local or remote SMTP input, all subsequent messages on the same
connection are queued." In other words, the behaviour was fully documented.

2. But I'm not always grumpy: I have implemented a new option called
queue_only_load_latch, which defaults TRUE. If you set it FALSE, the load is
re-evaluated for each message in a connection, which should give you the
behaviour that you want on very long connections. This patch is committed to
CVS, and so should be in tonight's snapshot.

--
Configure bugmail: http://www.exim.org/bugzilla/userprefs.cgi?tab=email

--
## List details at http://www.exim.org/mailman/listinfo/exim-dev Exim details at http://www.exim.org/ ##