2009-03-13 05:06 AM
We need to snapmirror 1Tb of data onto a live filer at a remote site. We have a spare DS14 shelf, all our filers are FAS270's or 2020's. All our filers are on 22.214.171.124 or higher. Both the current and destination FAS2020's are without external disk shelves. Is is possible to add this shelf to the source FAS2020, create a new aggregate, volume snap the data onto it. Halt the filer, remove the shelf, ship it and add it to the FAS2020 at the remote site and snapmirror update.
Would something along these lines work.
Should both the source and destination filers be using software based disk ownership. Any pointers on this would be great.
2009-03-13 11:57 AM
What is the connectivity on the DS14 shelf? Has it got full ESH modules (with GBICs for connectivity)? If so, you should be able to do what you are asking. All filers will want to be using software based ownership, which should already be the case given your system models. You want to make sure you are always SnapMirroring to a destination that is either on a version of OnTap higher or the same level as the source. You won't be successful if it is a lower OnTap level.
2009-03-14 08:24 AM
I'm a bit weak on the software based disk ownership.
Does the data on the disk get wiped if you change the disk ownership ??
Should the disk ownership on the 14 disks be changed when the shelf is connected to the destination filer
or the source filer ??
2009-03-14 01:37 PM
Software ownership just simply controls who owns the disks themselves. Their membership of aggregates / raid groups and the data stored on them is totally independent and transparent to this.So long as you just change the ownership, all this will stay totally intact (which is why the filers are easy to work with!). When you move the disks around, the new filer will offline the aggregate on the disks as it will be marked as foreign, just online it and you are good to go.
You can use disk remove_ownership to free the disks up before you move them, this is an advanced command (and as ever, use with caution!).
filer> priv set advanced
filer*> disk remove_ownership <disk_name>
filer*> priv set
Replace <disk_list> with the actual disks you want to remove ownership of (e.g. 0a.16, 0a.17 and so on...). I believe you have to do this one disk at a time, but I haven't got a system to try this on for you at the moment I'm afraid.
When you move the disks across, they will either be owned automatically, or you just use "disk assign all".
Hope this makes sense? Give me a shout if you got any more questions.