Over the weekend, immediately after getting the 1755260 update, spamd
went into kind of a tail spin in which it would no longer spawn new
children.  After about 90 minutes, all other children are killed,
leaving spamd running but not actually able to process anything.

Googling indicates it has something to do with the "select returned
error on server filehandle" error, but I'm not finding much other than
that (like the actual cause or a solution).  It seems like it has
something to with the sa-update, but whether it's specific to the
update in question or just that an update took place isn't obvious.

Did anyone else see an issue what that particular update?

This is on Ubuntu running 3.4.1 (3.4.1a-0b~trusty to be exact), with
spamass-milter calling spamd (with default children settings).

Log entries around the sa-update show (the first line is pre update):
Aug  6 06:52:39 wanamaker spamd[1334]: prefork: sysread(20) not ready,
wait max 300.0 secs
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: new lowest idle kid:
none
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: select timed out (via
alarm)
Aug  6 06:53:34 wanamaker spamd[17580]: Use of uninitialized value
$selerr in concatenation (.) or string at
/usr/share/perl5/Mail/SpamAssassin/SpamdForkScaling.pm line 332.
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: select returned error
on server filehandle:
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: no spare children to
accept, waiting for one to complete
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: child closed
connection
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: child 1334: entering
state 1
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: new lowest idle kid:
1334
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: child reports idle
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: overloaded,
immediately telling kid to accept
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: ordered 1334 to
accept
Aug  6 06:53:34 wanamaker spamd[1334]: spamd: accept() on fd 5
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: sysread(19) not
ready, wait max 300.0 secs
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: child 1334: entering
state 2
Aug  6 06:53:34 wanamaker spamd[17580]: prefork: new lowest idle kid:
none



--
Public key #7BBC68D9 at            |                 Shane Williams
http://pgp.mit.edu/                |      System Admin - UT CompSci
=----------------------------------+-------------------------------
All syllogisms contain three lines |              sha...@shanew.net
Therefore this is not a syllogism  | www.ischool.utexas.edu/~shanew

Reply via email to