View previous topic :: View next topic
|
Author |
Message |
dean
New User
Joined: 26 Jan 2010 Posts: 2 Location: albany, ny
|
|
|
|
I need to migrate from FDR/ABR to DFSMS/HSM. Does anyone know where there is any good documentation for this procedure? |
|
Back to top |
|
|
expat
Global Moderator
Joined: 14 Mar 2007 Posts: 8797 Location: Welsh Wales
|
|
|
|
IBM website might help, but this is a project I done myself some time ago, so will try and dig up my notes etc. etc. |
|
Back to top |
|
|
expat
Global Moderator
Joined: 14 Mar 2007 Posts: 8797 Location: Welsh Wales
|
|
|
|
What I recall about moving FDR/ABR to DFSMShsm.
Once the sysprogs and storage team, (me), have the system set up to use both FDR/ABR and DFSMShsm, with FDR/ABR in passive mode. i.e. will only Recall / Restore datasets, the fun begins.
The tricky bit was moving the backup versions as obviously you could not restore a backup with the same name while the original file still exists.
a) Recall the original file from FDR/ABR archive.
b) Rename the original file. (and migrate off if large)
c) Restore file from the oldest FDR/ABR backup.
d) Perform HSM backup of file.
e) Delete file.
f) Perform steps (c) through (e) for progressively newer backups.
g) When the backups migration is complete, rename file back to original name and migrate to HSM.
h) Move on to the next file.
I did an analysis of the expiration dates of all of the files held on FDR/ABR and started with the files that had the longest retention as this allowed the files with shorter retention periods to expire naturally without any intervention required as the longest retention files were slowly migrated to HSM, with files newly created being handled by HSM as a matter of course.
The whole migration process from FDR/ABR to HSM took about six to nine months. Good luck |
|
Back to top |
|
|
dean
New User
Joined: 26 Jan 2010 Posts: 2 Location: albany, ny
|
|
|
|
All our FDR backups will expire before the project is complete and all the new backups are being created by HSM so we won't have to migrate any backups. Also all our new migrations ar being done by HSM. Maybe I'm simplifying the process but isn't it as easy as recalling the migrated datasets and re-migrating them?
Thanks for your help. |
|
Back to top |
|
|
expat
Global Moderator
Joined: 14 Mar 2007 Posts: 8797 Location: Welsh Wales
|
|
|
|
It is |
|
Back to top |
|
|
Pete Wilson
Active Member
Joined: 31 Dec 2009 Posts: 581 Location: London
|
|
|
|
One thing to note with moving to DFHSM is it is not as flexible for scheduling because it is not Batch and doesn't allow processing to be as selective.
You can have a window for Backup or Migration on each LPAR in the Plex when these processes run, and a maxtask value in each LPAR (up to 15 max I think). The Storgrp settings then determine if they'll be processed by whatever task in the windows.
This requires some care because DFHSM is VERY CPU hungry and can be prone to not completing tasks within the windows if you have very large storgrps. It can take some time to get it tuned.
Maybe one day IBM will allow a separate BCR and MCR records for each Storgrp so they can run different schedules, but I'm not holding my breath!
The Tape Recycle is also very CPU hungry and can take forever. |
|
Back to top |
|
|
|