(this is a SUN Only Product, as far as I know)
Requires a second network card in each server, and a fake IP address.
Some of the Problems we have had:
The software points requests to the fake IP to the ACTIVE server. This
doesn't work well thru some firewall configurations because the Server
replies with its IP address and the firewall won't let it back thru.
Once failover from Primary to backup occurs, you must manually switch back
to the Primary once it is operational. The system doesn't make the Backup
the primary and does not switch back if it fails.
The system automatically tries to restart any process that fail, can be a
pain when you actually want to kill a process for maintenance reasons.
Good things :
Users in the middle of a transaction when the failover occurs will get a
timeout or two if they apply during this time. Once everything is up on the
secondary, the transaction can be applied with no loss of data.
We have not yet had the Primary fail on its own and the Secondary take over.
We have failed the Primary on purpose to test the failover, so we don't know
how everything will work in a real world failure, but the testing has
worked.
Hope this gives you a place to start.....
-----Original Message-----
From: Action Request System discussion list(ARSList)
[mailto:ARSLIST@LISTSERV.VISTAIT.COM] On Behalf Of John Neels
Sent: Thursday, December 30, 1999 8:35 AM
To: ARSLIST@LISTSERV.VISTAIT.COM
Subject: Failover systems for Remedy and Oracle 8.0.5
<< File: John Neels.vcf >> Has anyone implemented a seamless failover
between a two Remedy servers and
two Oracle Servers. If you could elaborate it would be great. We need
seamless failover for a 24x7 shop
John Neels, M.B.A., P.M.P.
Project Manager/Systems Engineer
Cypress Systems
230-5151 x118
johnn@cypress-systems.com
<>
To Unsubscribe send e-mail to:
body of text is: unsubscribe arslist
Previous Topic
Index