| Author |
Message
|
| ivanachukapawn |
 Posted: Tue Jul 29, 2008 7:13 am Post subject: Establishing a VIP |
 |
|
Knight
Joined: 27 Oct 2003
Posts: 561
|
| I need to setup a VIP for a SUSE Linux server and use it on the sender channel at a remote company's site. When this Linux server fails we would plan to reconfigure the VIP to point to a backup server at another location. I know that this sort of scheme is best solved with HACMP or Veritas but in this case the architect has elected to handle the failovers manually. My question is: How (using what software/utility) can this VIP be created? And in a failover scenario, how can the VIP be modified to point to the backup server? |
|
| Back to top |
|
 |
| zpat |
| Posted: Tue Jul 29, 2008 7:50 am Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
| Why don't you use IP names and change the DNS entry when needed to refer to a different IP address? |
|
| Back to top |
|
|
| ivanachukapawn |
| Posted: Tue Jul 29, 2008 8:05 am Post subject: |
|
|
Knight
Joined: 27 Oct 2003
Posts: 561
|
| The problem is that I am a network ignoramus. Do companies usually run their own DNS server? In either case is DNS reconfiguration typically responsive enough to be useful in a failover scenario? |
|
| Back to top |
|
|
| zpat |
| Posted: Tue Jul 29, 2008 11:20 pm Post subject: |
|
|
Jedi Council
Joined: 19 May 2001
Posts: 5866
Location: UK
|
Internal IP names use internal DNS server. It's not instant to update but is often used for DR purposes.
Or just change the IP address at the other end to the DR address. |
|
| Back to top |
|
|
| PaperHouse |
| Posted: Thu Jul 31, 2008 1:30 am Post subject: Establishing a VIP |
|
|
Newbie
Joined: 18 Jul 2006
Posts: 8
Location: London
|
Assuming for a second that there is a Firewall between you and the remote company!
You could perform the change at the Firewall by putting in a NAT (Network Address Translation). This changes the underlying IP address for an external connection.
As an aside, at our company we deem manual changes in Disaster scenarios as non-compliant. If you think about it, in most disaster cases you are not going to get any time or resources to make ANY manual changes. We have Architects and PM's who try it on all the time, but they get sternly rebuffed. If I were an external client to your organisation I would be horrified if you told me that there wasn't an automatic failover in the event of server failure. |
|
| Back to top |
|
|
| ivanachukapawn |
| Posted: Thu Jul 31, 2008 5:13 am Post subject: |
|
|
Knight
Joined: 27 Oct 2003
Posts: 561
|
Paperhouse,
I agree with you wholeheartedly. That is why I have always implemented automatic recovery via Veritas. In this case the client has no requirement for continuous availability. Rather, they have a job they want to run every night for a couple of hours. If the server fails, they want to restart this job on another server but they want the application to connect with the same queue manager name and IP. |
|
| Back to top |
|
|
| sami.stormrage |
| Posted: Thu Jul 31, 2008 9:42 am Post subject: |
|
|
Disciple
Joined: 25 Jun 2008
Posts: 186
Location: Bangalore/Singapore
|
Then think the solution calls for clustering rather than a HA implementation, since its a batch job and more over the window for failure is small.. Active support could be the key in maintaining SLA's. And implimenting HA solution for a pretty small thing (the only thing) is a costly affair.
_________________
*forgetting everything  * |
|
| Back to top |
|
|
|
|
