This post is more than 5 years old
1 Rookie
•
10 Posts
0
1745
February 13th, 2014 11:00
TimeFinder questions about Snap
Hi everyone,
I'm actually working in an environnement fully configured with TDEV (Thin Provisionning). My customer is taking Snaps of over 250 luns using actually a Clariion CX-4 to then mount the Snap onto a BK server and save the data on the Snap onto tapes.
Now, they have a VMAX40K and want to replicate what they are doing with the Clariion in the VMAX. I would like to test this so i could help the customer, but this is pretty much the first time i'm working with TimeFinder and am not too sure from where to start. So, from the described situation, can anyone give me advice ?
Thank you,
Christian
No Events found!
dynamox
9 Legend
•
20.4K Posts
0
February 13th, 2014 12:00
don't use the same host for this, typically you take snapshot of production volumes and present them to a backup server. Because these snapshot volumes will be identical, you don't want to mount them on the same host as production, it will confuse the host (logical volume manager). There are procedures to change volume group header/disk signature to be able to mount snapshot volumes on the same host as production volumes ..but you need to figure out what that process is and test test test first, you could bring production system down if you are not careful.
For your test if you only have one host to play with, activate snapshots, remove SRC devices and only then present TGT devices. This will allow you to validate data is there.
dynamox
9 Legend
•
20.4K Posts
0
February 13th, 2014 11:00
on VMAX you have two options, traditional Timefinder/SNAP with VDEVs or the new Timefinder VP/SNAP. (see attached document)
1 Attachment
docu40346_White-Paper_-Implementing-TimeFinder-VP-SNAP-for-Local-Replication.pdf
carriech82
1 Rookie
•
10 Posts
0
February 13th, 2014 12:00
Hi Dynamox,
Thank you for the document. I'm wondering how it should be prepared at host level. I guess, i need to create a partition on the source lun and place data on it, then do the procedure described in the document and snap the SRC volume.
For testing purpose, i have the TGT volume assigned to the same host as the source, can this cause an issue ?
I created a Device Group and placed the SRC and TGT volume in it, then made the pair and activate the snap (VP/snap). The GUI said after 3 minutes that the copy was done, but i couldn't see the data on the TGT volume. Not sure why.
Thank you,
Christian
carriech82
1 Rookie
•
10 Posts
0
February 13th, 2014 12:00
Thank Dynamox,
It makes perfect sense. I will try that, i have another host i can use to test, it should be more representative that way. I may have complimentary questions for you once i test it out.
Thank you for your help so far,
Christian
dynamox
9 Legend
•
20.4K Posts
0
February 13th, 2014 13:00
sure, have fun
carriech82
1 Rookie
•
10 Posts
0
February 14th, 2014 07:00
Hi Dynamox,
On a VMAX fully configured in Thin Provisionning, is the VP/SNAP with TGT volume using a Thin pool the only option available ? Are the SAVE device used with VP/SNAP ?
Can you use VDEV and SAVE device when the SRC devices are TDEVs ?
Thank you,
Christian
Quincy561
1.3K Posts
0
February 14th, 2014 08:00
I'm not sure why you would want to use VDEV/SAV Snaps if you already have a VP pool you can use. I would guess you don't have space on the drives to create a proper SAV pool, plus the added complexity that adds.
carriech82
1 Rookie
•
10 Posts
0
February 14th, 2014 08:00
Hi Dynamox,
Ok, so it mean that it is possible to mix SRC Thin Device with VDEV as TGT (Using a SAVE pool) ? In this case, would it be still SNAP/VP or the tradional TF/Snap that should be used ?
Also, in order to use a VDEV as a Snapshot TGT device, is there something special to do with the volume other than create it. IThe process of pairing will still need to be done.
Thank you,
Christian
dynamox
9 Legend
•
20.4K Posts
0
February 14th, 2014 08:00
Yes.it uses space from the pool. If you want to use vdevs you will need to create a dedicated save pool
carriech82
1 Rookie
•
10 Posts
0
February 14th, 2014 09:00
Hi Quincy,
I'm still unsure of the process behind TF, probably why my questions may sound a bit strange.
Actually, it is more in testing than anything. The customer has a pool (default) with like 10 TB of SAVE devices in it. Do you know if when using VP/SNAP, the VMAX use the SAVE device or not ?
Thank you,
Christian
Quincy561
1.3K Posts
0
February 14th, 2014 09:00
SAV devices are for use with Snap with VDEVs.
It is important with VDEV/SAV Snaps that the SAV pool be properly configured to handle the write load on the source devices when the Snap will be activated. The SAV pool will see a 100% write workload when activated (from any writes to the source that need to be preserved). If the SAV pool can't keep up, it can cause performance issues on the whole array.