Just in the planning stages on this one...
We have 4 ESX hosts all connected to an HP EVA. So when we implement VCB, 1) where do the snapshots get written during the backup? and 2) how much disk space does it take up?
Does it take up an amount equal to the current size? I'm trying to figure out the space requirements because we don't have a lot of free space left on the EVA. Thanks!
I too use the HP EVA. So far so good.
Snapshots will be created on the same VMFS volume that the VMDK is located on.
The rule for space is a tough one, but figure it shouldn't really ever get to be more than 20% or so AS LONG AS you take care of them and don't let them sit around too long. If you neglect them, they will grow and grow but will never exceed the size of the disk itself. The best approach to determine the size is to monitor the sizes of the snapshots and know what is typical for your environment.
Good Luck!
I too use the HP EVA. So far so good.
Snapshots will be created on the same VMFS volume that the VMDK is located on.
The rule for space is a tough one, but figure it shouldn't really ever get to be more than 20% or so AS LONG AS you take care of them and don't let them sit around too long. If you neglect them, they will grow and grow but will never exceed the size of the disk itself. The best approach to determine the size is to monitor the sizes of the snapshots and know what is typical for your environment.
Good Luck!
It depends:
Full-VM backup will freeze the machine, open a redo file, copy chunck of 2 Gigabyte on your backup server till the size of Full-VM.vmdk is reached, then commit the redo file and finally start to write to your tape (or whatelse...). This will get a little time (well, depending on full vm size), however little compared to the second option (the following)
Full/Incremental/differential backup (file level) that will open a redo file, mount the freezed snapshot of vmdk on VCB Proxy server, do a file level backup of the mounted image (writing to tape directly, except if you have a staging area), unmount and commit the redo log.
In both cases the redo log will grow proportionally with the time that VM remains freezed.
In my opinion you have to start with 12 - 15 % of free space on LUNs and then fine tune the free space,
hth
\aleph0
\aleph0
____________________________
###############
If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!
are you happy with the info?
\aleph0
____________________________
###############
If you found this information useful, please consider awarding points for "Correct" or "Helpful". Thanks!!!
Yes, thank you everyone. That was the information I was looking for.
