I am running two "clusters" using carp for network failover.
I use rsync every 15 minutes for the simple webapp which issues
x509 certs. A script runs on each node to check if it is
master if so it makes a crl, if not it pulls the directory hierarchy
from the master.
The other cluster does the same for the web pages, but uses Mysql
replication to keep the databases in sync.
Sooo ho hoo muuuuch cheaper than our AIX HACMP clusters on EMC .....
80-90% of the functionality for ~5% of the cost.
Seems to me that there was/is some daemon on the redhated step child of
an distro that you could use to look for changes in an file or dir
structure that you could use. I'll see if I can rember/find it, I though
it was from SGI.....
This may or may not help
http://oss.sgi.com/projects/fam/links.html
Jens Mayer wrote:
Dear all,
we are thinking about a scenario on how to set up a server offering http, ftp
and a few postfix/mailman driven mailinglists with a redundant failover. I'm
_not_ talking about load balancing here - only the master is serving, while
the slave sits still and waits, probably with all services shut down until
taking over.
While the networking part can be handled by carp, I'm collecting ideas on how
to keep the local file systems in synch - especially for ftp users and the
mailinglist archives. The synchronization will be done via a dedicated cross
coonect cable directly between the boxes.
I've seen nice concepts like DRBD (www.drbd.org), offering a "RAID-1" network
block device, but did not find anything like that for OpenBSD.
Of course there's always the possibility of scripting something own using
rsync and friends, but I'm curious if some of you have a similar setup
running and can share some ideas, thoughts and big red warnings.
Kind regards,
Jens
