Hello,
I am in the process of upgrading SA from version 2.63 to 3.1.8.
I have done it on a test machine for a few mail adresses including
mine and I came across several problems. I am running spamd and
spamc is invoqued from procmail in /etc/procmailrc (system-wide
invocation)
1) spamd stopped analyzing mails after some time. I found the following
message in /var/log/maillog :
spamd[1958]: prefork: select returned -1! recovering: Bad file
descriptor
I found this problem is described here :
http://issues.apache.org/SpamAssassin/show_bug.cgi?id=4590
http://issues.apache.org/SpamAssassin/show_bug.cgi?id=5313
http://issues.apache.org/SpamAssassin/show_bug.cgi?id=5422
I applied the patch from 5313 by downloading an nightly build
from CVS, extracting the two modules : SubProcBackChannel.pm
and SpamdForkScaling.pm and putting them in place.
=> this problem seems to be cured and I hope version 3.2.0 will
not show the problem.
2) Since problem 1 is fixed, I observed at least once that I had an
extra spamd child owned by a user and that this spamd child were
busy with the ~/.spamassassin/bayes_toks.
A lockfile ~/.spamassassin/bayes_lock was in place with 2 identical
lines containing the PID of this spamd child.
A message could be found in /var/log/message :
spamd[17207]: bayes: bayes db version 2 is not able to be used,
aborting! at
/usr/lib/perl5/site_perl/5.8.0/Mail/SpamAssassin/BayesStore/DBM.pm line
196, <GEN239> line 954.
This spamdd process was apparently stuck because it stayed for
hours. Then following an action that I forgot, the lockfile
disapeared but a ~/.spamassassin/.nfs000xxxxxxx appeared.
Only when I manually killed the spamd child, this .nfsxxxx
file vanished.
I know the format of the bayes DB has changed but as I understood,
the conversion is supposed to be performed by SA itself when
opening the DB R/W. It does not seem to work well.
=> Any experiences of a migration of the bayes DB files when
upgrading from 2.6x to 3.1.x ?
=> What is the best way to deal with this ?
Thank you all.
Jean-Michel BARBET.
--
------------------------------------------------------------------------
Jean-michel BARBET | Tel: +33 (0)2 51 85 84 86
Laboratoire SUBATECH Nantes France | Fax: +33 (0)2 51 85 84 79
CNRS-IN2P3/Ecole des Mines/Universite | E-Mail: [EMAIL PROTECTED]
------------------------------------------------------------------------
