Drive restore suggestions pls

[RajeshR]

Hello,
We have a cluster shared drives on windows with 2 hosts, one of the drive requires restore to different drive within same cluster.
This drive has 4TB of data with 1 million files approx and shared widely in the network.
We have came across these scenarios many times.
Normally we are taking down time and restoring the source drive data to new drive, before start of restoring we are disabling sharing for not to access/update/add any files to make sure perfect restore.

Since this server is running with LAN backup/restore operations, i can't make it lanfree. (no free ports for zoning)

my question is how can i reduce my restore time.

If i restore this drive to new drive from latest backup without downtime and then stopping share of the source drive(once restore completed) .. then performing incremental backup for source drive and then restoring incremental using " -ifnewer " option is that works? is this option only restores only newly added files and modified files from latest incremental backup?
since i can't use the replace all option, what it will do for modified files during latest incremental backup.?? because the file names will be same only last modified and last backedup time and size changes.

any suggestions please..

TSM server 7.1.6
TSM Client 7.1.6
and i am using 3 mount points and 5 resources and 10g transacation byte limit.

 

[moon-buddy]

If this was a UNIX box, I would use the DD utility.

Have you tried Ghost or a similar utility? I have not touched Ghost for a long time but I believe there is a later version that allows disk mirroring even if the target disk is much larger than the source disk.

Since this is a TSM forum, have you tried Image backup and restore?

 

[pmello]

Hello!

You can restore only the directory structure first and after this, only the files. For this, use the options DIRSONLY and FILESONLY in restore command.

Take a look on this articles for improve your next restore:

Dirmc
http://www-01.ibm.com/support/knowl....ibm.itsm.client.doc/r_opt_dirmc.html?lang=en

Using collocation with tape drives
http://www-01.ibm.com/support/knowl...f.doc/c_optimization_collocation.html?lang=en

 

[marclant]

If i restore this drive to new drive from latest backup without downtime and then stopping share of the source drive(once restore completed) .. then performing incremental backup for source drive and then restoring incremental using " -ifnewer " option is that works? is this option only restores only newly added files and modified files from latest incremental backup?
since i can't use the replace all option, what it will do for modified files during latest incremental backup.?? because the file names will be same only last modified and last backedup time and size changes.

That plan is your best bet. Keep in mind that when you do your 2nd restore with "-ifnewer -replace=yes", it will have to go through all the files backed up previously on that filespace to determine if they need to be restored or skipped. So that may take a while.

An image backup like moon-buddy suggested could be quicker to recover. You could do image+incremental. Restore the image and then restore the incremental: https://www.ibm.com/support/knowledgecenter/SSGSG7_7.1.6/perf/r_client_bup_method_imagebackup.html

 

[RajeshR]

Thank you all, For now we have to do it within specified time, so i have used maximum sessions.
For next restore will plan it with either ifnewer or image backup to restore.

 

[RajeshR]

Can't image backup take more than one session (i mean data transfer sessions to tape) despite having maxnummp=5 resourceutilization=5
Trying to backup one shared drive using image backup, But backup running on only 1 data transfer session to tape.
If i try normal backup it's taking only 2 drives despite having maxnummp=5 and resourceutilization=5.
We have also DIRMC set to DISKPOOL.

 

[moon-buddy]

It is an image - so the data stream is contiguous (serial in nature) which means it cannot be broken down in multiple segments.