Hi,
we have an Exim server configured to forward emails to another
mail system (as a spam filtering system with Maia Mailguard (Amavis
and Spamassassin)).
We have a pretty standard Debian configuration setup for Exim,
with the following alterations (These are most of the alterations
we have):
router/01_amavis:
callout:
debug_print = "Initiating recipient callout to my mail host..."
driver = manualroute
verify_only
transport = recipient_callout
route_list = * mail.mailgateway.no byname
host_find_failed = defer
amavis:
driver = manualroute
# Don't run (condition=0) if received from port 10025
# as message has already passed through amavis
condition = "${if eq {$interface_port}{10025} {0}{1}}"
transport = amavis
route_list = "* localhost byname"
self = send
no_more
[EOF]
acl/30_exim4-config_check_rcpt:
[... lots of config ... ]
accept
domains = +relay_to_domains
endpass
message = unknown user
verify = recipient/callout=10s
# At this point, the address has passed all the checks that have been
# configured, so we accept it unconditionally.
accept
[EOF]
In the Exim mainlog we can see these:
2009-08-06 11:08:23 Failed to get write lock for
/var/spool/exim4/db/callout.lockfile: timed out
On a regular basis, sometimes 10-15 seconds apart.
And lsof reveals this:
quicksilver:/var/spool# lsof|grep lockfile
exim4 6659 Debian-exim 5uW REG 8,1 0 180348
/var/spool/exim4/db/callout.lockfile
exim4 7018 Debian-exim 4u REG 8,1 0 180348
/var/spool/exim4/db/callout.lockfile
exim4 7019 Debian-exim 4u REG 8,1 0 180348
/var/spool/exim4/db/callout.lockfile
When we ran lsof earlier, a lot more processes were also "attached" to
the lockfile,
but only one has the W (write lock for the entire file).
This is a testing environment with very few emails going through
(although spammers
have already managed to figure out it deals with mail..)
Any ideas?
TIA,
Morten
--
Morten W. Petersen
Manager
Nidelven IT Ltd
Phone: +47 45 44 00 69
Email: morten@???
