ADSM-L

[ADSM-L] 答复: [ADSM-L] 答复: [ADSM-L] backup stgpool - command issue

2012-05-18 21:18:33
Subject: [ADSM-L] 答复: [ADSM-L] 答复: [ADSM-L] backup stgpool - command issue
From: Victor Shum <victorshum AT CADEX.COM DOT HK>
To: ADSM-L AT VM.MARIST DOT EDU
Date: Sat, 19 May 2012 09:07:03 +0800
Dear Rick :

Thanks your advices.



Best regards,
Victor Shum

-----邮件原件-----
发件人: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] 代表 Rick
Adamson
发送时间: 2012年5月18日 21:14
收件人: ADSM-L AT VM.MARIST DOT EDU
主题: Re: [ADSM-L] 答复: [ADSM-L] backup stgpool - command issue

Victor,
Respectfully, I suggest you read the section of the TSM Server Admin or
Concepts guide to get a complete understanding how the server utilizes
primary tape volumes, and how disaster recovery manager (DRM) manages
copy-pools.

Basically, ask yourself the question; if you remove the primary pool tapes
from the library how is the TSM server going to be able to copy that data to
the copy-pool, and how is it going to reclaim the free space in the
copy-pool?

At a very high level here is how it should work.
1. The library contains all primary pool volumes and enough scratch tapes to
maintain an exact copy of your primary pool plus data base backups and tape
reclaims.
2. As new data is received it is stored in your primary pool then the
"backup stgpool" command mirrors the primary with the copy pool.
3. The "move drmedia" command ejects the newly generated copy pool tapes
from the library and alters their state so the TSM server know that they are
offsite and no longer in the library.
4. As they age data on the copy pool tapes expires and a process called
"reclaim" is run to rebuild the tape eliminating the free space left by
purged data. The newly rebuilt tape is then scheduled to go offsite and the
old tape is deleted from the storage pool so it can be reused as scratch
again.

Now, I have been in situations where the TSM Server's data outgrows the tape
library and will not accommodate all tapes. When this happens, tape
management becomes a hands-on manual process. Understand that when the TSM
Server needs data for any purpose, such as a restore or a reclaim, it uses
the primary pool copy  of that information, if it is not in the library the
server will display a prompt on the console to request it. If it is not
supplied the process will fail and will mark that tape volume as
"unavailable". When this happens the only way to make the tape available
again is to issue a command to alter the state of the volume back to a
status of readwrite.

While the solution to your situation is not complicated, it is not easily
resolved. Simply stated you either need to increase the size of your library
or reduce the retention so that all primary pool, and some scratch tapes can
be accommodated. 

Having someone monitor and hand-feed tapes to the server on an ongoing basis
is probably less cost effective than the other options after you consider
the cost of the persons wages over the long haul.

Again, this is just a high level explanation and I stress again that you
spend an hour or so to read the section in the manual that details tape
handling, especially as it pertains to DRM. The information is free and it
will be time well spent.

You can find it here:
http://www.ibm.com/developerworks/wikis/display/tivolidoccentral/Tivoli+Stor
age+Manager+Extended+Edition 

Good Luck !


~Rick


-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Victor Shum
Sent: Thursday, May 17, 2012 8:43 PM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: [ADSM-L] 答复: [ADSM-L] backup stgpool - command issue

Dear Rick :

We did not use the DRM now, 


Since the existing tape library slot (48 slot)is not enough to hold all
backup of primary pool; 14 version of user data (each version require 4 to 6
LTO5).  So we try to implement the backup like;


1.  daily schedule backup data to primary pool (daily selective / full
backup)
2.  backup stgpool to copypool
3.  backup TSM DB
4.  checkout all tape in the library
5.  check in new tape in the library
6.  checking any expired tape recycle the daily job

the copypool' tape will be transport to DR site, while the primary pool tape
will be store at the safe and reuse for after 14 day

We think we just daily duplicate the primary pool to copypool (using backup
stgpool), and check out all active tape.  Manually trace when the tape shall
be expire; then reuse (cycle) those tape again to make the things work...

So if we need to have the data retention period of 14 but don't have enough
tape slot in the tape library to hold all primary pool tape media.  We have
no way out !!

OR Shall we just shorter the retention period like: from 14 to 10 day, so
that the tape library can hold all the tape inside.  Use DRM to generate and
manage offsite tape form us ??

Best regards,
Victor Shum

-----邮件原件-----
发件人: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] 代表 Rick
Adamson
发送时间: 2012年5月17日 21:10
收件人: ADSM-L AT VM.MARIST DOT EDU
主题: Re: [ADSM-L] backup stgpool - command issue

Victor,

Assuming it is a copy pool volume (since primary storage volumes should
always be available in your library).
 
Investigate the volume status on your TSM server:
Use the "query drmedia" command determine the volume's current state, if in
fact it should be out of the library it's state would be either "offsite,
courier, vault, etc.". This tells TSM that the volume should not be in the
library and do not attempt to mount it. 
If the volume's state is "mountable" then the TSM server thinks the volume
is in the library. Use the "move drmedia" command to apply the correct
status to the volume. 

Additionally, check the access status of the volume by using the query
volume command to make sure it is still readwrite, and not readonly or
unavailable, as the server could have altered its access status when it was
unable to use it.

For a complete explanation of the DRM process refer to the admin guide.

~Rick
Jax, FL.


-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
Victor Shum
Sent: Thursday, May 17, 2012 7:13 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: [ADSM-L] backup stgpool - command issue

Dear all :



Due to tape library slot limitation, we implement TSM 6.3 with the following
scenarios :



Hardware

TS-3200 with 4 x LTO5, 48 slots



TSM configure

3 x primary pool ( A, B, C) for 1 x copy pool (for offsite - off_data)

4 cycle x 7 day of tape media

Daily backup : 4 ~ 8TB





Workflow

1.       Different hosts running different backup job, so that all the
tape
Drive will be occupy during the daily backup.

2.       At the end of day, we use backup stgpool to make a copy of
three
primary pool

3.       Check out all tape (current cycle)

4.       Insert a new set of tape

5.       Audit tape library, check in tape..etc

6.       Repeat 1



But we found the backup stgpool cannot be completed last-night.  From the
actlog, it keep report : ANR1228I Removable volume AAA208L5 is required for
storage pool backup. (SESSION: 10245, PROCESS: 199).  While this media was
used for backup in previous cycle (2 weeks ago).



So I am thinking is some of the files has been deleted on the hosts side
yesterday, so the existing tape (inside the tape library) does not content
those files.  But the "backup stgpool" has to copy all "active" files to the
copy-pool; that is why TSM try to mount other tape has the files on it??
Or I just hit a bug of TSM ??





Best regards,

Victor Shum