• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 8717
  • Last Modified:

Adding new scratch tapes to TSM library

Hi All,

I have inserted a new tape in library with label 007162 and I want to add it to scratch pool. I have searched on google and got following commands, but none of the commands are working, giving following output :

label libv L3494B search=bulk checkin=scratch labelsource=barcode
ANR8494E LABEL LIBVOLUME: An option specified is not valid for 349X libraries

checkin libvolume L3494B search=yes status=scratch checklabel=barcode
ANR8403E CHECKIN LIBVOLUME: Operation not allowed for 349X libraries

Now the problem is the tape has been moved to some slot in library, how can I add that tape to scratch pool, I have other 9 tapes which I have not loaded in the library yet, as I need to confirm first the commands are working. kindly provide me with the commands so that I can add the new tapes as scratch in library. TSM version is 5.5 running on AIX 5.3. Its bit urgent.

Thanks
virgo
0
virgo0880
Asked:
virgo0880
  • 7
  • 6
1 Solution
 
woolmilkporcCommented:
Hi,

1) SEARCH=BULK and LABELSOURCE are not allowed for 3494 libraries

LABEL LIBVOLUME L3494B SEARCH=YES CHECKIN=SCRATCH

2) CHECKLABEL=BARCODE is not allowed for 3494 libraries and you must specify a device type, most probably 3592 ("Jaguar") in your case.

CHECKIN LIBVOLUME L3494B SEARCH=YES STATUS=SCRATCH CHECKLABEL=YES DEVTYPE=3592

Both commands will work.

 (1) is for unlabeled tapes, but will work for labeled tapes as well. Add OVERWRITE=NO to preserve existing labels.

Note: All newly inserted cartridges will be checked in at once using the above commands.

Should you want to checkin only selected tapes use 007162 (i.e. the volume name) instead of SEARCH=YES in either of the above commands.

wmp
0
 
virgo0880Author Commented:
I have done this and it is showing me the volume as scratch in library. Now, when I tried to put remaining 8 volumes and executed the command, I am seeing that it is kicking out the tapes back to the IO port, what may be the issue...is it that I need to clean the label of the tapes.

Thanks
Virgo
0
 
woolmilkporcCommented:
Issue Q ACTLOG in dsmadmc to find the error messages.

Could it be that the carts only have an external paper label but no volume label on the magnetic medium itself? In this case use LABEL ..., not CHECKIN ..., and use OVERWRITE=YES.

But first, what's in Q ACTLOG?

wmp
0
VIDEO: THE CONCERTO CLOUD FOR HEALTHCARE

Modern healthcare requires a modern cloud. View this brief video to understand how the Concerto Cloud for Healthcare can help your organization.

 
woolmilkporcCommented:
... or maybe there is no free drive of the indicated device type?
0
 
virgo0880Author Commented:
No, the drives are free and also I cleaned two tapes and re-inserted the same, but it did the same thing. In the activity log it is showing errors :

03/23/11 14:31:50     ANR8807W Could not write label 006142 on the volume in    
                       drive TDRV2 (/dev/rmt12) of library L3494B because volume
                       is already labeled 006142. (SESSION: 10073, PROCESS: 353)


What can be done in this case.

Thanks
virgo
0
 
woolmilkporcCommented:
Use
CHECKIN
instead of
LABEL
or specify
OVERWRITE=YES
0
 
virgo0880Author Commented:
Yes, that worked. Now all of  my tapes are check in as scratch tapes. Just a question, my TSM is eating a lot of scratch tape, what can be done in this case and how can I find out out how the scratch utilization has increased and who is using that ?

Thanks
virgo
0
 
woolmilkporcCommented:
Do you use COLLOCATE=NODE or even COLLOCATE=FILESPACE in your tape stgpool definitions?

Check with

SELECT STGPOOL_NAME, COLLOCATE, NUMSCRATCHUSED from STGPOOLS order by NUMSCRATCHUSED desc

COLLOCATE eats up many tapes, because each node's data or even each filespace's data are placed on distinct volumes.

Check for the stgpools with the most used scratch tapes which node would use them up:

Q NODEDATA * STGPOOL=stgpool_name

Last, check if there are volumes with very poor utilization.

SELECT  VOLUME_NAME, STGPOOL_NAME, PCT_UTILIZED from VOLUMES where PCT_UTILIZED<2 order by PCT_UTILIZED

If there are lots this could be a hint that you're running too many backup jobs against the same stgpool in parallel.

If all the above seems normal you just have many data to backup.

wmp



0
 
virgo0880Author Commented:
I have taken the output of the commands given by you, can you have a look at it and provide me some recommendations on that so that I can improve something on this. The output is attached in a file.

Thanks
virgo tsm-output.txt
0
 
woolmilkporcCommented:
You have several of very poorly utilized tapes, particularly in TBACK11 which is due to COLLOCATE=NODE defined for that pool.
Do you need collocation? If not, change to COLLOCATE=NO (UPD STGP ...).

Most of your tapes are in CBACK5/6 and TBACK6, however, and I don't see lots of nearly empty tapes there.
You could repeat the utilization query with PCT_UTILIZED<5 or so, to get more output.

And please take care not to run all your backup jobs against a given storage pool at once. Define more schedules, to spread the jobs over a longer period of time, or choose a higher " Schedule Randomization Percentage".

wmp


0
 
virgo0880Author Commented:
Ok, what is the use of collocation and how it works. Is it advisable and advantageous. Also, if my tapes are less utilized they will be definitely used in next backup schedules ..right ?

Schedule randomization % is set to 10 currently, is there any recommended value for that.

Thanks
virgo
0
 
woolmilkporcCommented:
COLLOCATE specifies that the server should attempt to keep data belonging to a
single client node, group of client nodes, or client file space stored on as few volumes as possible.

In case of node recovery the number of mounts will be minimized this way.

It is advantageous only if your recovery window is very narrow and if you can afford the additional tape consumption.

Yes, your tapes will be uitilized next backup.

The needed amount of "filling" volumes in a stgpool is based on the number of tape volumes requested in parallel in a given time period and the number of available tape drives in that same period, whichever is lower.

Randomization is expressed in percent of the configured schedule duration, which is 1 hr. by default.
So with randomization 10% the start times of the backup jobs triggered by a given schedule will get scattered across a time window of 6 minutes by default.

If there are some schedules triggering many jobs I'd recommend increasing their duration to at least 5 hours (or more, depending on the number of jobs).
If all your schedules would trigger many jobs I'd recommend leaving the duration untouched and increasing randomization to at least 50% instead (or more, see above).
0
 
virgo0880Author Commented:
Will check the same.

Thanks
virgo
0

Featured Post

Independent Software Vendors: We Want Your Opinion

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

  • 7
  • 6
Tackle projects and never again get stuck behind a technical roadblock.
Join Now