Huwa
asked on
Exchange Disaster recovery switch on a FE/SMTP server
Hi all,
We are planning to move our exchange servers (2003 sp2) on to new hardware (ESX Virtual servers). The back end servers we plan to install as normal and move the mailboxes over. The FE i am thinking of running the restore switch.
What I would like to do to have as little downtime as possible on our frontend server which is also our main smtp server. The server is running on a physical dell server, Exchange 2003 SP2.
What I would like to do is
1. Setup a new virtual server (ESX).(same os,service packs and fixes as physical server)
2.Disconnect the network cables from the physical server, reset physical server account in our domain.
3.On the new virtual server rename it to be the same as the physical server and give it the same IPs ( 2 smtps, default for normal mail & 1 for tls)
4.Install exchange on new virual server with disasterrecovery switch.(SP2 with same switch)
5 Put cert back on FE server and am hoping after this that all works as normal.
6.Format old server.
Am I been to ambitious to believe this will work with little problems ?
Why this way, and not just install a fresh server ? this will I believe save me alot of time no reconfiguration of firewalls, the thwarte certificate we can continue to use and not have to purchase a new one and most importantly if i understand the switch command properly it should pull all the settings for the exchange from AD which I am guessing less chance of anything going wrong as the setup we have now works very well.
My question is has anyone any experience using the switch on a FE exchange server that is also acting as a SMTP. Is there any gotchas, any thing i should watch out for.
Thanks in advance
Hugh
(Yes Exchange 2007 is nice :-) but will wait for the next version)
We are planning to move our exchange servers (2003 sp2) on to new hardware (ESX Virtual servers). The back end servers we plan to install as normal and move the mailboxes over. The FE i am thinking of running the restore switch.
What I would like to do to have as little downtime as possible on our frontend server which is also our main smtp server. The server is running on a physical dell server, Exchange 2003 SP2.
What I would like to do is
1. Setup a new virtual server (ESX).(same os,service packs and fixes as physical server)
2.Disconnect the network cables from the physical server, reset physical server account in our domain.
3.On the new virtual server rename it to be the same as the physical server and give it the same IPs ( 2 smtps, default for normal mail & 1 for tls)
4.Install exchange on new virual server with disasterrecovery switch.(SP2 with same switch)
5 Put cert back on FE server and am hoping after this that all works as normal.
6.Format old server.
Am I been to ambitious to believe this will work with little problems ?
Why this way, and not just install a fresh server ? this will I believe save me alot of time no reconfiguration of firewalls, the thwarte certificate we can continue to use and not have to purchase a new one and most importantly if i understand the switch command properly it should pull all the settings for the exchange from AD which I am guessing less chance of anything going wrong as the setup we have now works very well.
My question is has anyone any experience using the switch on a FE exchange server that is also acting as a SMTP. Is there any gotchas, any thing i should watch out for.
Thanks in advance
Hugh
(Yes Exchange 2007 is nice :-) but will wait for the next version)
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
Hi Guys
Thanks for the replies I can agreee that the setting up of the server itself doesnt take long but when you look at the changes that need to be made on
1. Mx records.
2.ISP to change pointer for reverse lookup.
3 SPF records
4 Firewall changes for new IP addresses
The way I have described above i am guessing 1 - 2 hours downtime and no changes needed on the 4 points above that i have mentioned.
You mentioned replication Metasha, we have 2 back end servers would these not replicate public folders and smtp setting when the server is back up and running.
Thanks
Hugh
Thanks for the replies I can agreee that the setting up of the server itself doesnt take long but when you look at the changes that need to be made on
1. Mx records.
2.ISP to change pointer for reverse lookup.
3 SPF records
4 Firewall changes for new IP addresses
The way I have described above i am guessing 1 - 2 hours downtime and no changes needed on the 4 points above that i have mentioned.
You mentioned replication Metasha, we have 2 back end servers would these not replicate public folders and smtp setting when the server is back up and running.
Thanks
Hugh
Huh?
1. Mx records.
2.ISP to change pointer for reverse lookup.
3 SPF records
4 Firewall changes for new IP addresses
Why do you have to make any of those changes?
Just use the same IP address as the old server, or switch things around.
I can replace a frontend server in less than two hours. I don't do anything with DNS. The firewall change takes two, three minutes?
You don't replicate the public folders to a frontend server.
SMTP Server settings aren't stored in either the domain or Exchange. They are unique to the server.
-M
1. Mx records.
2.ISP to change pointer for reverse lookup.
3 SPF records
4 Firewall changes for new IP addresses
Why do you have to make any of those changes?
Just use the same IP address as the old server, or switch things around.
I can replace a frontend server in less than two hours. I don't do anything with DNS. The firewall change takes two, three minutes?
You don't replicate the public folders to a frontend server.
SMTP Server settings aren't stored in either the domain or Exchange. They are unique to the server.
-M
ASKER
OK I see what your getting at basically set up a new server set up new IPs and new name, remove old cleanly. make the new installation a FE, and give it same IPs as old server.
But i still believe for our routing groups, bridge head servers between sites, exchange policies regarding replications of PF are all kept in AD.
What i didnt know and what you pointed out is the SMTPs would need to be set up newly, thanks for that info.
Hugh
But i still believe for our routing groups, bridge head servers between sites, exchange policies regarding replications of PF are all kept in AD.
What i didnt know and what you pointed out is the SMTPs would need to be set up newly, thanks for that info.
Hugh
RGs, policies etc are stored in the Exchange org, but those would apply to a new build or a DR build. You gain little from doing a DR install.
-M
-M
if you going for a new build, you would definately need to add the new servername in the bridgehead for your SNTP connector & RGC
ASKER
Just for Info, all SMTP settings (smarthost, IP addresses, sender ID IMF, bridgehead etc were already on the new server after disaster recovery. I am not 100% sure if this is done using the disaster recovery switch dut I am pretty sure it is which was my objective. Agreed a pure OWA server a disaster recovery or new install would make no difference. but when it is hosting smtp services (as we had default plus TLS/smtp) then didsaster recovery is a better way to go.
Thanks for the comments.
Hugh
Thanks for the comments.
Hugh
The settings to come out of the domain for an Exchange 2003 frontend server is negligible. The bulk of the settings are on the server itself which would need to replicate. The time saving would be minutes doing it this way, if that.
I would just build a new server with a new name, configure it identically, then move the SSL certificate across. That will allow you to confirm that both servers are configured identically.
-M