This post is more than 5 years old
75 Posts
0
2424
Need to migrate File Systems to new Luns
Hello,
probably this is an already threaten thread....
I've a NS480 box with NAS 5.6.49-3 version.
At the moment, my clar5_performance pool is composed of 4x 4+1 450 GB disks in RAID 5 configuration (6,3 TB of datas).
In one month, I'll receive brand new 600 GB disks (probably 20), and I want to migrate on them all of my cifs file systems. (consider that I'll configure all of disks in RAID 5 configuration [4+1 disks and creating 2x luns per storage group... ].
How should I do them? with Clariion Lun migration? With File system copy?
Can I do it without service outage?
Thanks for any answer.
PS: I'm using AVM
Riccardo Barone
Rainer_EMC
8.6K Posts
0
November 22nd, 2010 12:00
that does require a Replicator license though
riker82
75 Posts
0
November 22nd, 2010 23:00
OT:
that's probably what I've look for archive old data in slower disks (SATA) or different Celerra....
What about DR restore?
In my company, we'll do cold restart in case of disaster, restoring via ndmp (using a Datadomain appliance in VTL mode) all of our cifs datas in a different Celerra located in a DR site (in housing in HP Italy).
What If I'll use FMA (archiving old files in SATA disks in the same production Celerra)?
Can I backup this old datas once (via ndmp) and restoring them in case of disaster?
What about stub files in a non-archived file systems?
thanks a lot for your help
riker82
75 Posts
0
November 23rd, 2010 00:00
Thanks Rainer,
Yes I've got Replicator V2 license and I'll probably use It.
thanks a lot to everybody for your help!
riker82
75 Posts
0
November 23rd, 2010 00:00
Hi Rainer,
probably Lun migration is the less disruptive solution and:
1) Yes I'll use the same RAID criteria
2) Yes I'll use the same number of disks
3) NO, I'll not create the exact LUN SIZE
about the third point, I'll waste so much space... that's why I cannot use Lun migration
Rainer_EMC
8.6K Posts
0
November 23rd, 2010 01:00
no problem there - just create two LUNs per RG of the same size as the old ones and two with the remainder that you can later use to extend the old fs or for new ones
no wasted space there
Rainer_EMC
8.6K Posts
0
November 23rd, 2010 05:00
Well – backup and DR restore typically depends on your DR planning and what / where your archive target is
You could use NDMP backup with a setting that backups the archive target just like there was not stub (basically then NDMP does a passthrough like a client would).
That’s certainly feasible for a “fast” archive target like another Celerra share on the same or a different Celerra.
In this case the amount and time for a full backup stays roughly the same.
Other customers just use a NDMP backup that backups just the live data and the stubs and then either:
- Don’t backup the archive at all – since its on a different and maybe replicated system
- Backup the archive – but on a less aggressive schedule since archived files never change
Or they archive to platform that has builtin replication and self-healing like Centera or Atmos that typically never get backed up.
That has the advantage that your regular full backup will be smaller and faster.
Since you already have a Data Domain another option would be to archive to that DD in file mode – there the adavantage is that typically any archived files you send to the DD are already stored there from previous backups so the archive costs you very little space (as long as the backup holding the file isn’t expired and purged)
Hope that helps
Rainer
riker82
75 Posts
0
November 23rd, 2010 07:00
Unluckly my Datadomain (DD565) is running @ very low space in this period (and I don't like its performance for cifs purpose)
Since our business decided that our DR plan involves a COLD RESTART (restoring 48 hr before datas) we probably need to backup production and archived data and the same frequency.... Since Celerra replicator permits only to have a sync as a time of 24 hrs before.
Rainer_EMC
8.6K Posts
0
November 23rd, 2010 10:00
I think archiving to a DD would be done via NFS and not CIFS
Also the workload to the archive store is very different from the workload of a production file system – its mostly sequential write and read of decent-sized files
If you need to restore state of 48 hours before a DR – why don’t you setup checkpoints on the DR replication system for that ?
Then you could revert that DR systems file systems to that two days ago checkpoints and then copy that back to a repaired/replaced production system
Rainer
riker82
75 Posts
0
November 23rd, 2010 23:00
That's a very good idea. I could propose it to our business (hoping that I don't need so much disk space in order to keep this checkpoint....)
THANKS A LOT RAINER!!!!