Standby Continuous Replication (SCR)
This post has been moved to
http://www.exchange-genie.com/2007/08/standby-continuous-replication-scr/
Monday, August 27, 2007
Subscribe to:
Post Comments (Atom)
This post has been moved to
http://www.exchange-genie.com/2007/08/standby-continuous-replication-scr/
Posts
33 comments:
Nice work, Brian!
Excellent work Brian - this is very useful information! :-)
Matt Ward
I had some problems setting up SCR in my lab. The database did not seed on the target machine. The workaround I came up with was Suspend-StorageGroupCopy SGName -StandbyMachine Target, delete all log files from Target, Update-StorageGroupCopy SGName -StandbyMachine Target (from Target), and Resume-StorageGroupCopy SGName -StandbyMachine Target
Gr8 work Brian.
Have a question for you.
What will happen to already connected users. I am sure outlook with go in Offline line mode. What do we need to do from client/end user prospective to get them connected automatically. Changing manually the server name, its a difficult thing on every end user desktops.
Thanks
Rahul
rahulxarora,
The answer really depends upon the client that you are using, OWA and Outlook 2007 (as long as autodiscover is working) will work with out any problem however legacy like Outlook 2003 will have problems only if the original server is no longer online.
rahulxarora,
I forgot to mention, if you use the /serverrecover switch you can use the same server name to removes any client configuration changes. I am currently working on my next article related to that SCR option.
Very useful information.
About how long does it take to run the Set-Mailboxdatabase SCRTARGET\RECOVER\RECOVERDB -AllowFileRestore:$true
and the
Mount-Database SCRTarget\RECOVER\RECOVERDB
commands?
If a primary email server dies, then what kind of time frame is realistic for getting a passive SCR machine up and running? 30 Min?
I would say there are a number of factors that will dictate that answer.
Which method would you use database portability, /recoverserver, or /recovercms each has its steps?
Also I have just recieved new code and want to rerun my tests. A fix has been deployed so that eseutil may not have to be run.
I would say you could script many of the manual tasks to speed up the process and if you have large db that require eseutil that will take time.
I dont want to be vague but the time it will take will vary for each person.
Hi! I've been looking for really useful info on the web, but I could rarely find the ones which are useful... luckily i found your blog about Exchange SCR! This is really fantastic!!!
I hope you'll share more info as good as this in your upcoming blogs :)
Gerald Elizalde
Is it possible to use SCR on a combined HTS/Mailbox server?
How long should it take to seed a 80gb storage group over a 100mb wan link?
I've successfully seeded a couple of my sg's that are all they same size and it literally took 2-3 days. Is that normal????
How long should it take to seed an 80gb storage group ober a 100mb wan link?
I've seeded a couple of my SG's that are all about the same size, and they literally took 2 or 3 days to finish???? Is this normal?
I dont believe there are any issues with combo roles, in beta one node was a combo and worked
Willis,
I would say no that is not normal however the are a number of factors mainly bandwith. You have a 100mb link but how saturated is it?
How taxed are the boxes currently?
Hi!I have an issue after enabling SCR which is my Database will not seeded automatically and through error such as "The Database not found after log replay Source:MSExchangeRepl event ID:2106"...Plz help me....
This is some great information; thanks for the effort!
I have setup SCR and it seems to be working correctly; going to test the changeover this evening from our production CCR mailbox server to the SCR Target in a different subnet & AD site.
Hopefully this is an easy question, but I cannot seem to find the answer aynwhere:
I have a huge amount of log files replicated from the CMS to the SCR target. I know that once the SCR target has 50, it replays these into the DB, and the SCR Db is the same size as the production Db.
My question (finally) is: HOw do I make the SCR Target purge those log files it's already used?
Brillant Post thank you,
I have set SCR up now and have it seeded to the DR site, but when I use get-stroagegroupcopy status -standbymachine etc, it says not configured.
Just set it up and left defaults for relay time so may have to wait 24 hours to see action on logs etc, but this has me a touch concerned thoughts?
Which machine are you running the command from? Also the log copy should start as soon as you have everything setup so the target should be recieving logs.
I was running the cmd from the target machine. Ran it again same machine still same answer not configured. No logs as of yet. It seeded and created the log folder,
and mail database folder / file .
Ran the enable-storage etc cmd from the target as well originally.
Lost as to why it's not replicating. thoughts going forward? THx Greg
Ok, i Have been setting it up on 2 pairs of servers. I ran the suspend, then update on one pair and now the logs are there. Then ran the get status on the source and got healthy. So all good there. Other pair gave me no valid replication string message, and appeared not to have scr configured despite seeding etc. so I have disabled and manually deleted seeded folders files on target server, but when I run the enable it tells me the logs already exist etc. running enable on target server. Love ex2K7 but SCR is getting me down
Update, have 1st pair working now.
Was able to re-seed second pair, suspend, then update, but if I go to source machine and run the get status cmd it says not configured. Also there are no logs in the target log folder.
Greg
Got it! For some reason dispite having SCR disabled it would recreate the database, and log folder. However I broke down and renamed the source storage group and database. then moved to new folder paths etc. Ran through enable , seed process and all good now. Thanks,
Greg
Can you have multiple sources on one target? If so, what are the drawbacks?
Yes, you can have multiple source sever -> one target..... draw back would depend, you need to have the same paths on sources and target this would limite you to database portability unless you went through a bucnh of hoops.
Should the log file prefix of the source storage group and the target storage group should be same.
I have successfully activated target database on scr server but the log file prefix was different. Because i have activated the sixth storage group database and on the target scr already there were two storage group.
How can we change the log file prefix of a storage group?
Do you have to have LCR enabled before enabling SCR?
My Exchange is saying LCR is not enabled so it cannot enable the storagecopy.
Nice work.
One question, i have two exchange 2K7 server with two databases. Is it possible for each server to be source and target of each other ?
N.B
Yes you can do that and use database portability to bring them online. I guess one question is are these both local or in 2 offices. If local you may want to think of using CCR which would be much better for you.
Thanks heaps for this article!
I've just run through all the steps successfully, but it fails right at the last hurdle.
When I try to mount the DB I get an Active Directory error. The server I am mounting it on is a DC/GC, however the source server was the PDC and holds all the FSMO roles - do I need to transfer the roles for the DB to mount? Tried rebooting and leaving it unmounted for a period of time with no luck - and after the reboot the Information Store service won't even start...
Now, when I stop all the exchange services on the source server and re-enable the NIC - the services all start and the DB mounts. This suggests to me that the target server NEEDS to have the PDC online for the recover to work. If this is true, I'm a little disappointed as I thought SCR was meant to be an easy way to replicate exchange and recover quickly.
Really appreciate any comment on the situation.
Cheers,
Matt
http://mattlog.net
great work..very insightful.
quick question.
how do we go back to the original server?
-remove any trace of the original SG and DB by deleting the SG and DB? and setup SCR again?
I guess the answer would depend on the failure however yes you can just clean up the server and get everything functional, a simple move mailbox would work to bring the users back you could also do the process in reverse. There is a technet article that walks you through
Brian,
Can you explain what the /a switch does when running eseutil with a lossy failover? I can't seem to find it documented.
this article has a good break down
http://www.computerperformance.co.uk/exchange2007/exchange2007_eseutil.htm
Post a Comment