Mailing List Archive

rsyslog clustering
Hi All,

I have a question about rsyslog configuration in a clustered environment. Currently we have two data centres and I was going to deploy a syslog server at both of these locations. I was hoping also implement a "master" log server here in the office where we could easily view the logs from both locations in one place. However, looking at the schema for Syslogs.SystemEvents, it's using ID as the primary key which is an auto_increment integer. Obviously, this won't work. I'm trying to avoid any code change as well.

I've noticed that syslog-ng have a commercial solution with the concept of a "relay" which basically concatenates the messages in the data centres and feeds into a master server. Is it possible to configure rsyslog to behave in a similar fashion?

Failing that, I was thinking of using MySQL database replication whereby each data centre would have a different database name and could then replicate their databases to the master server. However, is it possible to specify a failover MySQL server in the rsyslogd configuration?

I'd be very interested to hear from anyone who's implemented a similar solution. Thanks for your time!

Regards,
Leigh
rsyslog clustering [ In reply to ]
Hi,

> Hi All,
>
> I have a question about rsyslog configuration in a clustered
> environment. Currently we have two data centres and I was going to
> deploy a syslog server at both of these locations. I was hoping
> also implement a "master" log server here in the office where we
> could easily view the logs from both locations in one place.
> However, looking at the schema for Syslogs.SystemEvents, it's using
> ID as the primary key which is an auto_increment integer. Obviously,
> this won't work. I'm trying to avoid any code change as well.
>
> I've noticed that syslog-ng have a commercial solution with the
> concept of a "relay" which basically concatenates the messages in
> the data centres and feeds into a master server. Is it possible to
> configure rsyslog to behave in a similar fashion?
>
> Failing that, I was thinking of using MySQL database replication
> whereby each data centre would have a different database name and
> could then replicate their databases to the master server. However,
> is it possible to specify a failover MySQL server in the rsyslogd configuration?

Yes, I don't have the links to this on-hand, but if you check the mail
archives you'll find others who have replied to me with the links to explain
how to implement a backup MySQL database if the primary one was to fail.

I personally haven't implemented this (yet) as the MySQL environment is
already HA'ed, which lessons the need for me to have a failover MySQL
environment (although I intend to implement it at some stage in the future
just for my own interest).

Regards,

Michael.

> I'd be very interested to hear from anyone who's implemented a
> similar solution. Thanks for your time!
>
> Regards,
> Leigh
> _______________________________________________
> rsyslog mailing list
> http://lists.adiscon.net/mailman/listinfo/rsyslog
------- End of Original Message -------