I have two servers, currently running Red Hat Enterprise (EL3) with 6
hard drives each 2 for OS (mirror) 2 for data (mirror) and 2 for Mysql
DB (mirror)
both are identical.
now without spending any extra cash on hardware, is there any way to
get a replicated or clustering filesystem?
i've looked high and low, and all over google, sourceforge, freshmeat,
slashdot, etc and I haven't had any luck.
Mysql replication doesn't work because it only fails-over and only
manually. no fail-back.
mysql clustering is OK, except the memory of the systems is only 1 gig,
and our database will grow to 1 gig in under a month. so that's out.
LinuxHA may be a possibility but it seems overly ***bersome and not
really workable for such a small environment.
NFS doesn't work, sure manual replication, but now fail over/back
ability at all.
the only possible things I've found which might work is CODA or
inter-mezzo, but they seem experimental.
yes I know the obvious answer is to use a SAN or storage array, but as
I said, we have no budget for hardware.
The systems have 3 NICs each, so some out of band syncing is certainly
possible.
I briefly considered NetwareFS and LDAP, neither are feasible.
anyone got any bright ideas? for something automated?
thanks
PS - this email address is dead, don't bother trying to contact
there...
|
