A
A
Andrey Osty2012-01-24 12:07:10
Clusters
Andrey Osty, 2012-01-24 12:07:10

In ms sq 2008 r evolution to standard cluster

The end of the ms sq 2008 r evolution period has come. We bought a standart license and it turns out that it is not possible to upgrade evolution to standart in cluster systems!!! We have already inflicted severe bodily harm on the system architect. Advise what to do, the server has already become combat, and that it is part of the Real-time system. It is highly undesirable to stop, as there will be data loss.

Answer the question

In order to leave comments, you need to log in

2 answer(s)
P
PavelN, 2012-01-24
@PavelN

1. As far as I remember, Standard supports clustering in 2 nodes
2. There will be a stop anyway.
3. I see the solution as follows:
* install a new instance on the same servers, create a cluster
* step by step transfer the settings of the master database, msdb can be transferred with a backup, if there were changes in the model database - also with a backup
* transfer the databases to the Full Recovery model mode if they are in some other
place * back up the necessary databases and restore them on a new instance in norecovery mode (this is necessary to minimize the downtime)
* put out the solution
* make a differential database backup or a transaction log backup. Restore it on a new instance.
* put out the old database server instance
* reconfigure the paths from your solution to the new server
* start the solution
in this case, the downtime will be small. For 15-20 minutes of downtime, you can roll a backup copy of the log, change paths and confuse everything.
4. You can also use mirroring to minimize downtime and auto-switch, but this needs to be tested first.

D
dbmaster, 2012-01-26
@dbmaster

As I understand it, you want to use iron too.
I have never done this, and here a person writes that it is possible to do "Upgrade to a Different Edition"
blogs.technet.com/b/sqlman/archive/2010/03/16/sql-server-2008-upgrading-to-standard- enterprise-edition-from-evaluation-edition.aspx
If this option doesn't work, then I favor option #4.
Put another cluster in parallel and use "mirroring". Free up one disk in the current cluster and use it in the new one.

Didn't find what you were looking for?

Ask your question

Ask a Question

731 491 924 answers to any question