09-17-2018 03:55 AM
We have EV 12.0 and NetBackup 8.1.
We had our vault store partitions for mailbox and journal archives on a Centara, but now we are migrating to a Isilon.
I have to backup the vault stores now and I'm having slow throughput <500kb\sec
I reading that I should keep the vault stores size 400-500 GB for keeping backup windows short.
How do I tell the size of a partition? I can run a vault Store Usage Report, but that does not give me partition size.
Do I Create Roll Over partitons for my mailbox archiving and journaling archiving partitions?
Is there a limit to number of partitions that can be created?
Thanks
Pat
09-17-2018 04:53 AM
Hi,
I assume you use 2 Isilons/Centera's? with replication?
Why backup? I use Centera's for journal data, no backups there.
09-17-2018 05:16 AM
We are retiring the Centera and using the Isilon now.
09-17-2018 05:43 AM
Hello, again,
Let me rephrase :-). Both devices are in general used to perform replication to an identical device. Using a Centera has benefits where the EV Centera API informs EV items are secure. Isilon does not have that. (I know, we're in testing phase). The Isilon in DC1 is replicated to an Isilon in DC2. This replication is fast, and (so I am told) secure. Having to backup TB's of data, where you know the data is replicated, is (sometimes) overkill.
I 'solved' it by setting the Vault Store (and it's partitions) that use Isilon in backup mode, write a triggerfile (IgnoreArchiveBitTrigger.txt), then take the VS out of backup mode. EV will then consider all files created before the date/time of the triggerfile as secure, and update the databases accordingly.
If your backup is slow, it might have several reason. You should involve your backup solution vendor, and perhaps EMC, to assist you with that.
Does that better clarify?
09-17-2018 05:46 AM
Thank you, much appricated.
Pat
09-18-2018 05:58 AM
Question, what if there is corrupt data on EV, wouldn't it just be replicating that data? where would you get clean data from?
09-20-2018 12:03 AM
Hello again,
In short, **bleep** happens. You cannot prevent everything.
If you rely on replication of data, and data is corrupt, then yes, corrupt data will be replicated.
What if you notice after a few months, and you keep backups for 30 days? No way you are going to be able to retrieve that data from EV again. Will you then restore a backup of Exchange (if you keep that for a long time)? If you want to be really secure, you will need to talk to a partner/consultant/Veritas itself, to see how they advise to mitigate that risk.
09-21-2018 06:01 AM
How are you creating the trigger file? Is this a daily task for you?
09-24-2018 11:54 AM - edited 09-24-2018 10:49 PM
Hi again.
I use 2 batch files. 1 sets backup mode to the Vault Store(s) managed by a server, 2 clears backup mode for those stores.
in 1, I have the below command added.
echo > \\isilon\folder\vaultstorepartition\ IgnoreArchiveBitTrigger.txt
IGNORE. ECHO > works. I believe I have echo 2> but not sure. you need to test that. After backup is cleared, the txt extention is changed to .old, so you can write a new txt file the next run. Works like a charm.
09-25-2018 03:21 AM
Thanks again for the response.
10-01-2018 07:29 AM
You're welcome. If you one of the answer solves the issue, can you mark one of the entries as solution? Thanks