We run UV on a 2-node linux HA cluster.  The backup instance of UV is a
cold spare, though.  Failover is automatic, and less than 2 minutes, but
any active UV processes are terminated since UV is started fresh on the
backup node.  This may or may not cause a problem for you depending on
how your application is written.  If your code makes use of commit and
rollback, users have to repeat the process they were in the middle of
but your data will still be consistent.  Another consideration is the
possibility of broken files when you do OS level replication.  You can
avoid this risk by implementing UV transaction logging (don't know about
UD) or by sizing all your file so they use zero overflow.  I opted for
the latter in our case because it wasn't that difficult given the size
of our database.

If you need your application to run continuously without any
interruption, you might check into Vmware's Vmotion.  We currently have
two Vmware ESX servers (though not for UV) using shared storage on a
NetApp SAN.  Right now we're only licensing the Vmware HA capability,
which automatically brings any machines that were running on one ESX
server up on the other in the event of a physical server failure.
Vmotion takes this a step further and migrates the running virtual
machines to the other server without stopping them.  I hear it works
well, but we haven't opted to license it yet due to the cost.  If UV or
UD were implemented this way, I think that should mitigate the
aforementioned broken file risk since you reduce the risk of a physical
server failure interrupting a write to overflow.

-John

-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Baker Hughes
Sent: Monday, July 26, 2010 6:38 AM
To: 'U2 Users List'
Subject: [U2] 24 X 7 MV systems

Hey y'all,

I'm interested in hearing from folks who are currently on, or have
worked with fault tolerant MV systems.

We'd like to host our Business Layer on the MV system and serve It to
our e-commerce portals, instead of re-coding our business rules first in
Basic, then in .Net   In order to get there though we must meet the
primary business requirement of zero downtime (not even 2 minutes to
manually switch).  We're not talking about different levels of Raid -
it's assumed the storage array is up and available.  If the MV system
has a hiccup of more than a few seconds it needs to hot failover to a
backup twin sister.

Is anyone doing this or something close to it?  When I worked in public
safety, Stratus sold such an automatic hot failover.  I'm sure the
EnRoute folks are doing something like this still.  Maybe Nick G. or
Margaret M. is listening in today.

Thanks,
-Baker



  ________________________________
This communication, its contents and any file attachments transmitted
with it are intended solely for the addressee(s) and may contain
confidential proprietary information.
Access by any other party without the express written permission of the
sender is STRICTLY PROHIBITED.
If you have received this communication in error you may not copy,
distribute or use the contents, attachments or information in any way.
Please destroy it and contact the sender.
_______________________________________________
U2-Users mailing list
[email protected]
http://listserver.u2ug.org/mailman/listinfo/u2-users
_______________________________________________
U2-Users mailing list
[email protected]
http://listserver.u2ug.org/mailman/listinfo/u2-users

Reply via email to