cancel
Showing results for 
Search instead for 
Did you mean: 

Global cluster replication and changing storage

Tomislav
Level 4
Partner Accredited
Hi!

I have Global cluster where I have on Primary site four node cluster and on the Disaster recovery site one node cluster which in case of some problems on Primary site can takeover all services from Primary site cluster. Between these two clusters data is replicated with VVR. We want to replace old storage on Primary site with new one without breaking VVR replication. Is something like this possible or not. Are there any "how to" documentation for that?

Thanks in advance,
Tomislav
2 REPLIES 2

Roger_Zimmerman
Level 4
Hi, Tomislav,

there are two ways to do this with only Storage Foundation mechanisms.

The first one would be to make on the primary side an additional mirror to your existing volumes living on the new storage device. After syncing to the "new primary side raid system" you can delete the mirror pieces on the "old primary side raid system". Quite simple. No Volume is changing name and the storage landscape is not altering in logical way.

The second possible solution is to make an additional secondary site for the Volume Replicator on the new raid system. Unfortunately to do this you need an additional server for the time of migration as well. Ah, and yes, an additional license for Storage Foundation, at least temporary. When this additional secondary site is synced you can promote this to the new primary and after a time of checking and testing you can eliminate the former original site what is now a secondary.

The problem of both of these scenarios is that you have to provide two raidsystems for a certain time. So an "in place change" of the raidsystem is not that easy. There are more than one stage of mounting and syncing and consuming energy and fibrechannel controllers ... And, much worse, is is certainly not to be done in one day (or night).

Maybe all this effort can be avoided with a short downtime (some hours, when all steps are very good planned and prepared). Taking the replication to a paused state, making a backup of the original data space, shutting down the old raidsystem, bringing the new one in place and providing the LUN's, restoring the data and taking care for the Volume Manager volumes are online again (thank there is this logical layer volume) and restarting the replication is a quite usable way, depending of the amount of data we talk about.

Whith all those ways you have advantages and misadvantages. But they all have one thing in common: you cannot change the number of volumes (so you cannot consolidate to a smaller number of LUN's) and therefore your logical structure of the storage keeps the same.

Ah, another thing they also have in common: planning is essential!

Hope those little thougts are helpful.
Roger

Tomislav
Level 4
Partner Accredited
Hi, Roger!

Thanks for your quick answer.

I think that first option is the one that I will do. It seems that first solution is the simplest way to do it. Second solution and the one with backup/restore it seems to be too much complicated for me.
Volume consolidation for now isn't problem because everything should stay as is.
If I have to pay attention for anything for first solution please let me know.

Regards,
Tomislav