02-11-2012 01:09 PM
Guys,
I have two EV servers (one each for Exchange and FSA). The backup product we are using is TSM.
The TSM backup runs in a batch, which means I dont know when does the backup for File Servers target ends/completes.
I have to create a script to reside on EV server, which can be called by TSM when TSM ends the File Servers' backup.
The script has to;
1 - Put the vault stores and index locations in backup mode
2 - Call the TSM schedular to initiate the job to backup EV vault stores and Index locations through TSM to tape
3 - reset the archive bit
4 - Clear the backup mode
Could you please assist?
Thanks
Solved! Go to Solution.
02-11-2012 08:44 PM
Typically TSM calls the EV Scripts. That is how customers I have worked with in the past do it.
TSM call the SetBackupMode.bat script, does the backup, and then calls ClearBackupMode.bat create the trigger file and to clear backup mode.
02-11-2012 01:23 PM
Give this technote a read. It explains how to use a trigger file for backup software that will not remove the attribute. It was create for EV 8 but is still valid for EV 9 and 10.
Article: TECH67559 | | | Created: 2009-01-05 | | | Updated: 2010-01-20 | | | Article URL http://www.symantec.com/docs/TECH67559 |
02-11-2012 03:03 PM
thanks - I had been through this before.
If someone already have a script to call TSM scheduler to initiate EV vault stores and Index locations, please assist.
Thanks
02-11-2012 08:44 PM
Typically TSM calls the EV Scripts. That is how customers I have worked with in the past do it.
TSM call the SetBackupMode.bat script, does the backup, and then calls ClearBackupMode.bat create the trigger file and to clear backup mode.
02-12-2012 02:29 AM
as far i know - TSM has a incremental forever approach. Have you ever considered the needed time to restore a whole VaultStore Partion?
Wouldn't it be nice, if you could get rid of the long backups and restores?
Our approach is to replicate instead of backing up. (as archived data is static data...)
You might want to read our whitepaper realated to that:
http://www.quadrotech-it.com/files/EVnearSync-WhitePaper.pdf
Best regards,
Peter
02-12-2012 08:35 AM
02-12-2012 10:36 AM
Peter - this involves extra cost, probably per TB, right?
Do you have a licensing model for me to go through please.
02-12-2012 10:52 AM
TSM runs the File Servers backup as part of a batch which runs over-night.
So, we dont know when does the File Servers backup completes; but we know when does batch ends. We only want to kick-off the EV archiving task once the file server backup is completed.
Since we are unaware of the completion time of the File Servers backup, should we;
first - kick-off the EV archiving task (during the day)?
second - kick-off EV server backup during the day?
02-13-2012 12:08 AM
on number of users... not on TB...
If you want a ballpark figure of the costs (it's not that expensive...) please drop me a pm.
Thanks,
Peter
02-13-2012 12:16 AM
... if you do it the right way (object based and application integrated).
Did you ever saw a customer backing up his centera's?
As archived Data is static data and will never change, a 2nd reliable copy is enough for DR Proposes. Why having the same static saveset's in different Backup Sets?
We can do the same thing with our software for ANY KIND OF STORAGE! and because it's integrated in EV's saftety copy mechanism you'll have a RPO of 0 (no dataloss possible) and a RTO of some sconds.
/Peter
02-13-2012 12:17 AM
replication!=backup -agreed
02-13-2012 04:00 AM
02-13-2012 04:13 AM
that's the case we have a "delayed deletion" on the replica... So you are able to restore within a defined period of time (default: 6 months).
If you need to restore a whole deleted archive, how would you do it from backup?
Restore to an 6 Month old version of the EVDB's? ;)
From my expierence (...and i'm dealing since 10 years with EV) i never saw customers backing up their centera environment...
But that's just my 2 cents...
02-13-2012 04:27 AM
02-13-2012 04:39 AM
that's exactly the reason we're not using storage based (block based) replication (aka. mirroring).
The source as well the target Disk are completly independant so no logical corruption can be replicated.