snapshot compatibility list : Local vs. Shared Disk

[Michal_Mikulik1]

Hello,

in the Snapshot Compatibility List (http://www.symantec.com/docs/TECH126901 ), there is a section "Disk". I see that it reports supported snapshot combinations which are independent on array vendor. But can somebody explain difference between Local and Shared Disk sections?

Because, for example, "Local Disk" section contains also Alternate Client methods, so the disk must also "Shared" in some way here.

Regadrs

Michal

[Nicolai]

Disk Type "Shared Disk" seem to be linked to "Data Mover: Third Party Copy" in the "Snapshot Method" column.

I have not seen/heard any one using "Third Party Copy"

[RiaanBadenhorst]

Hi,

Snapshot compat is a bit confusing (as you've noticed).

What is your rerquirement / environment / scenario?

[Michal_Mikulik1]

I want to generally report to customer about choices for online JFS backups (both traditional and FlashBackup) via snapshots.

He has array models which are not supported for this FS very widely, so I wanna to find out something about general "Disk" methods.

[Nicolai]

My recommendation would be to go for client side deuplication and accelerator before looking into array snapshots. All form for array based snapshot are complex to manage/configure and does not scale very good.

Only in extreme cases where RTO/RPO is more important than $$$ does make sense to look at array snapshots.

Disk method is usually Veritas Storage Foundation - this is a clever piece of software, but isn't cheap.

[RiaanBadenhorst]

Hi,

All those DISK methods actually refer to Storage Foundation (Infoscale now). You need the volume manager installed and then you can configure a mirror, etc. The type of mirror you use will determine if the snapshot can be mounted off host on an alternate client (if required).

Prices of SF have been reduced but its probably a bit of overkill to buy that in your scenario, and it might not solve the issue. If you're doing this because backups are slow, you're backing up lots of small files, then this wont help. Accelerator, as Nicolai mentioned, would be a better solution.

If you need off host processing to reduce the load on your production host, or to get the source data closer to the backup target you can look at snapshot options. This is usually done with large databases.