[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-ha-dev
Subject:    [Linux-ha-dev] Clustering/File system/Trade offs
From:       <rschapel () edtech ! mcc ! edu>
Date:       2003-04-28 13:45:53
[Download RAW message or body]

Greetings All: 3 Different choices

If any of you can help, and give advice, THANKS up front! Stories welcome!

  I am a bit new to the clustering world and looking for some advice.
Currently I am looking at setting up an active-active file server
solutions in which I have two servers connected to a shared SCSI storage
array, with a pair of Adaptec 39160's (*). Now I have a decision from
this point, I could mount all the space on one server, then NFS share it
to the other server, I could then use both servers for file service, and
have a nice fail over scenario. Is this method going to be beneficial?
Obviously the fail over solution is wonderful, but do you typically see
a noticeable performance increase in using an active-active, sharing the
same space over an NFS share? Am I just making this more complex then it
should be? My instincts tell me I don't want all my clients connecting
to the same server (active passive solution), and by just splitting that
login overhead the units will benefit as a whole (The units will be
running Samba/Webspace server).

 I could also look at sharing the SCSI bus between both units at the same
time. I could actually mount different segments of the SCSI array on both
machines at the same time. I could use something like EVMS, Create
private containers, and keep a set of reserve disks, and then just add
the disks as space becomes a issue. How does this compare in the
performance to a traditional Active-passive fail over and the above
Active-Active solution Via NFS share. This is the approach I really want
to go with, I just don't know if I'll be shooting my performance in the
foot, sharing that SCSI bus. I guess I don't really know enough about
SCSI, but I am hoping one of you has dealt with issue, and could give me
some guidance. Thanks for you time!


(*) 2 Dell 2650's and a PV 220 with 12 73 Gig Drives.
--Randy Schapel
Educational Systems Mott Community College


_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev@lists.community.tummy.com
http://lists.community.tummy.com/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
[prev in list] [next in list] [prev in thread] [next in thread]