Discussion:
[Sqlgrey-users] Database down issue
Bastien LAMOUREUX
2010-01-14 10:42:59 UTC
Permalink
Did a few tests with SQLGrey yesterday with following setup
Version: 1.6.8-1 stable release

Server1 : postfix/sqlgrey
Server2 : postfix/sqlgrey
Server3 : Mysql Database backend for sqlgrey


It look that sqlgrey is a wonderful replacement for postgrey as I need to share database between my 2MX servers.
The only problem I seems to encounter while doing tests is that if sqlgrey loose connexion to database backend, it seems to shut down with a segfault and even if connexion come back a few seconds later, it does not restart and I encounter this kind of messages.
Is there a way for sqlgrey to restart automatically after a database connect issue (maybe in development release?)
Is it safe circumvent this by adding a /etc/init.d/sqlgrey start in a cronjob running every 10 minutes?


Jan 14 10:50:14 PCACTI01 kernel: [85965.629178] sqlgrey[17606]: segfault at 4 ip b7a488a4 sp bfe645b0 error 4 in libmysqlclient.so.15.0.0[b79da000+1a4000]
Jan 14 10:50:15 PCACTI01 postfix/smtpd[17673]: warning: connect to 127.0.0.1:2501: Connection refused
Jan 14 10:50:15 PCACTI01 postfix/smtpd[17673]: warning: problem talking to server 127.0.0.1:2501: Connection refused
Jan 14 10:50:15 PCACTI01 postfix/smtpd[17673]: NOQUEUE: reject: RCPT from mymachine.mydomain.com[myIP]: 451 4.3.5 Server configuration problem; from=<***@yousaas.com> to=<***@trucmuche.com> proto=ESMTP helo=<MyHELO>

Thanks for help
Lionel Bouton
2010-01-14 12:12:17 UTC
Permalink
Post by Bastien LAMOUREUX
Did a few tests with SQLGrey yesterday with following setup
Version: 1.6.8-1 stable release
Server1 : postfix/sqlgrey
Server2 : postfix/sqlgrey
Server3 : Mysql Database backend for sqlgrey
It look that sqlgrey is a wonderful replacement for postgrey as I need
to share database between my 2MX servers.
The only problem I seems to encounter while doing tests is that if
sqlgrey loose connexion to database backend, it seems to shut down
with a segfault and even if connexion come back a few seconds later,
it does not restart and I encounter this kind of messages.
Is there a way for sqlgrey to restart automatically after a database
connect issue (maybe in development release?)
The 1.7 releases are quite stable but see below.
Post by Bastien LAMOUREUX
Is it safe circumvent this by adding a /etc/init.d/sqlgrey start in a
cronjob running every 10 minutes?
No. If you get the segfault now, there's no telling how frequent it can
become, see below.
Post by Bastien LAMOUREUX
segfault at 4 ip b7a488a4 sp bfe645b0 error 4 in
libmysqlclient.so.15.0.0[b79da000+1a4000]
This is a segfault in the MySQL client library. There's either a bug in
it (or one library it uses) or an hardware problem. SQLgrey can't do
anything about it : the system it runs on is unreliable, so it can't be
reliable itself. My advice : lookup segfault problems with
libmysqlclient and your OS on the web and try to find if anyone else
with the same problem could find a solution.

Regards,

Lionel
Bastien LAMOUREUX
2010-01-14 17:21:19 UTC
Permalink
Did same test with Postgresql instead of Mysql
It does not behave the same way. SQLGrey does not shut down with segfault, and just let all mails "in"
Seems perfect to me



De : Bastien LAMOUREUX
Envoyé : jeudi 14 janvier 2010 11:43
À : 'sqlgrey-***@lists.sourceforge.net'
Objet : Database down issue

Did a few tests with SQLGrey yesterday with following setup
Version: 1.6.8-1 stable release

Server1 : postfix/sqlgrey
Server2 : postfix/sqlgrey
Server3 : Mysql Database backend for sqlgrey


It look that sqlgrey is a wonderful replacement for postgrey as I need to share database between my 2MX servers.
The only problem I seems to encounter while doing tests is that if sqlgrey loose connexion to database backend, it seems to shut down with a segfault and even if connexion come back a few seconds later, it does not restart and I encounter this kind of messages.
Is there a way for sqlgrey to restart automatically after a database connect issue (maybe in development release?)
Is it safe circumvent this by adding a /etc/init.d/sqlgrey start in a cronjob running every 10 minutes?


Jan 14 10:50:14 PCACTI01 kernel: [85965.629178] sqlgrey[17606]: segfault at 4 ip b7a488a4 sp bfe645b0 error 4 in libmysqlclient.so.15.0.0[b79da000+1a4000]
Jan 14 10:50:15 PCACTI01 postfix/smtpd[17673]: warning: connect to 127.0.0.1:2501: Connection refused
Jan 14 10:50:15 PCACTI01 postfix/smtpd[17673]: warning: problem talking to server 127.0.0.1:2501: Connection refused
Jan 14 10:50:15 PCACTI01 postfix/smtpd[17673]: NOQUEUE: reject: RCPT from mymachine.mydomain.com[myIP]: 451 4.3.5 Server configuration problem; from=<***@yousaas.com> to=<***@trucmuche.com> proto=ESMTP helo=<MyHELO>

Thanks for help

Loading...