Hi all, i'm wondering whether any of you clustering experts could help.
Our apps team look after a trading system that uses an HP DL360 server
running SQL 2000 SP3a with a few small databases on it. This is a critical
server as it stores trading system information, market orders etc.
A requirement has now arisen to cluster this machine with another similar
server and connect each to an HP Smart Array, with the SQL databases being
installed on that.
Is this possible? What will happen to SQL or the OSs when the systems are
clustered and the storage array attached and configured? I assume i'll have
to back up the databases, cluster then attach the storage, then restore the
databases to the array.
Any information would be greatly appreciated!
Thanks
Andoni
Andoni
I think the easiest and safest way to accomplish this will be to reinstall
the SQL Server once the cluster is complete. Once the server is complete you
can attach your datafiles or to restore them from a valid backup. If your
application makes changes to the master database you should consider to name
your virtual server as the current production server (after renaming the
current server of course) this way you may be able to reuse your master
database.
Have in mind that all datafiles MUST be on the shared disk(s) so consider to
have enough space on the Shared storage, because the cluster Can't use local
disk for sql server datafiles, only for binaries.
Hope this helps.
Regards
Fernando Rivas
|||Hello Andoni,
You can do this it’s quite straight forward follow the steps bellow, but
make sure you read about and peice that is not clear on books on line.
1.Install clustering on new machine we’ll call it svr1.
2.Install clustering on old server we’ll call it svr2 and the sql on it sql1.
3.Once you have clustering working and failover of the resource groups is
working as expected with out error your ready to install sql server. Now if
you installed a default instance first time around on svr2 so that sql1 is
default and your apps need to connect to a default instance then this will
not work. In that case you let me know.
4.Install sql as a clustered resource we’ll call it sqlcluster
5.Apply service pack.
6.At this point you have your clustered instance and your non-clustered
instance running at the same time.
7.Make sure the cluster instance is on svr2 so that server has control over
the shared disk.
8.Then backup your sql1 instance.
9.Move the logons from sql1 to sqlcluster.
10.Detach the first database your interested in and then copy the data and
log files to the data and log shared drives you setup as part of the cluster.
11.Then attatch the same files to the sqlcluster instance the fix any
orphaned logins.
12.The run dbcc checkdb(databasename) on the database.
13.Rebuild indexes and update stats on the database.
14.Test that the application can connect to the new instance sqlinstance.
15.If so then repeat steps 10-14 for each database.
16.Once the migration is done I would disable the old server and not
de-install quite yet.
I made a lot of assumptions about your level of knowledge so it’s a part is
to vague then let me know.
Hope this helps
John Vandervliet MCSE, MCTS
"Andoni" wrote:
> Hi all, i'm wondering whether any of you clustering experts could help.
> Our apps team look after a trading system that uses an HP DL360 server
> running SQL 2000 SP3a with a few small databases on it. This is a critical
> server as it stores trading system information, market orders etc.
> A requirement has now arisen to cluster this machine with another similar
> server and connect each to an HP Smart Array, with the SQL databases being
> installed on that.
> Is this possible? What will happen to SQL or the OSs when the systems are
> clustered and the storage array attached and configured? I assume i'll have
> to back up the databases, cluster then attach the storage, then restore the
> databases to the array.
> Any information would be greatly appreciated!
> Thanks
> Andoni
No comments:
Post a Comment