TS3200 Mixed LTO3 and LTO5 mounting tapes to wrong drive

[CoMaboy]

Hello,

We've been running TSM 6.3.1.0 on Windows Server 2008 SP2 hooked up to a TS3200 tape library with two LTO3 drives. A couple of weeks ago we've added two LTO5 drives to the library and have been plagued by tape/drive mismatching problems since. LTO3 media is being mounted in the LTO5 drives and vice versa. Also, when a scratch volume is needed for the LTO5 pool, it tries to mount every LTO3 tape first before mounting the correct format and proceeding with whatever is was doing. All LTO3 scratch volumes get marked as private as a result.

This is what our setup looked like before adding the new drives and corresponding devclasses. One devclass LTO, one primary pool TAPEPOOL and one copy pool VAULTPOOL. We changed the names, adding _LTO3 to the stgpool names prior to creating the new ones.

Device Class Name: LTO
Device Access Strategy: Sequential
Storage Pool Count: 2
Device Type: LTO
Format: DRIVE
Est/Max Capacity (MB):
Mount Limit: DRIVES
Mount Wait (min): 60
Mount Retention (min): 2
Label Prefix: ADSM
Drive Letter:
Library: TS3200
Directory:
Server Name:
Retry Period:
Retry Interval:
Twosided:
Shared:
High-level Address:
Minimum Capacity:
WORM: No
Drive Encryption: Allow
Scaled Capacity:
Primary Allocation (MB):
Secondary Allocation (MB):
Compression:
Retention:
Protection:
Expiration Date:
Unit:
Logical Block Protection: No



Storage Pool Name: TAPEPOOL_LTO3
Storage Pool Type: Primary
Device Class Name: LTO
Estimated Capacity: 0,0 M
Space Trigger Util:
Pct Util: 0,0
Pct Migr: 0,0
Pct Logical: 0,0
High Mig Pct: 50
Low Mig Pct: 42
Migration Delay: 0
Migration Continue: Yes
Migration Processes: 1
Reclamation Processes: 1
Next Storage Pool: TAPEPOOL
Reclaim Storage Pool: TAPEPOOL
Maximum Size Threshold: No Limit
Access: Read/Write
Description: Primairy Pool.
Overflow Location: Overflow
Cache Migrated Files?:
Collocate?: Group
Reclamation Threshold: 100
Offsite Reclamation Limit:
Maximum Scratch Volumes Allowed: 75
Number of Scratch Volumes Used: 0
Delay Period for Volume Reuse: 0 Day(s)
Migration in Progress?: No
Amount Migrated (MB): 0,00
Elapsed Migration Time (seconds): 141
Reclamation in Progress?: No
Storage Pool Data Format: Native
Copy Storage Pool(s):
Active Data Pool(s):
Continue Copy on Error?: Yes
CRC Data: No
Reclamation Type: Threshold
Overwrite Data when Deleted:
Deduplicate Data?: No
Processes For Identifying Duplicates:
Duplicate Data Not Stored:
Auto-copy Mode: Client
Contains Data Deduplicated by Client?: No


Storage Pool Name: VAULTPOOL_LTO3
Storage Pool Type: Copy
Device Class Name: LTO
Estimated Capacity: 47.870 G
Space Trigger Util:
Pct Util: 24,4
Pct Migr:
Pct Logical: 99,0
High Mig Pct:
Low Mig Pct:
Migration Delay:
Migration Continue:
Migration Processes:
Reclamation Processes: 1
Next Storage Pool:
Reclaim Storage Pool:
Maximum Size Threshold:
Access: Read/Write
Description: Vault/Copy Pool.
Overflow Location:
Cache Migrated Files?:
Collocate?: No
Reclamation Threshold: 100
Offsite Reclamation Limit: No Limit
Maximum Scratch Volumes Allowed: 75
Number of Scratch Volumes Used: 49
Delay Period for Volume Reuse: 3 Day(s)
Migration in Progress?:
Amount Migrated (MB):
Elapsed Migration Time (seconds):
Reclamation in Progress?: No
Storage Pool Data Format: Native
Copy Storage Pool(s):
Active Data Pool(s):
Continue Copy on Error?:
CRC Data: No
Reclamation Type: Threshold
Overwrite Data when Deleted:
Deduplicate Data?: No
Processes For Identifying Duplicates:
Duplicate Data Not Stored:
Auto-copy Mode:
Contains Data Deduplicated by Client?: No

Then we added the new drives with the intention of using them for the on-site pool and using the old drives for the off site pool. So we created two new devclasses LTO5 and LTO3 along with the new primary and copy storage pools like so.

Device Class Name: LTO3
Device Access Strategy: Sequential
Storage Pool Count: 1
Device Type: LTO
Format: ULTRIUM3C
Est/Max Capacity (MB):
Mount Limit: 2
Mount Wait (min): 60
Mount Retention (min): 2
Label Prefix: ADSM
Drive Letter:
Library: TS3200
Directory:
Server Name:
Retry Period:
Retry Interval:
Twosided:
Shared:
High-level Address:
Minimum Capacity:
WORM: No
Drive Encryption: Allow
Scaled Capacity:
Primary Allocation (MB):
Secondary Allocation (MB):
Compression:
Retention:
Protection:
Expiration Date:
Unit:
Logical Block Protection: No

Device Class Name: LTO5
Device Access Strategy: Sequential
Storage Pool Count: 1
Device Type: LTO
Format: ULTRIUM5C
Est/Max Capacity (MB):
Mount Limit: 2
Mount Wait (min): 60
Mount Retention (min): 2
Label Prefix: ADSM
Drive Letter:
Library: TS3200
Directory:
Server Name:
Retry Period:
Retry Interval:
Twosided:
Shared:
High-level Address:
Minimum Capacity:
WORM: No
Drive Encryption: Allow
Scaled Capacity:
Primary Allocation (MB):
Secondary Allocation (MB):
Compression:
Retention:
Protection:
Expiration Date:
Unit:
Logical Block Protection: No


Storage Pool Name: TAPEPOOL
Storage Pool Type: Primary
Device Class Name: LTO5
Estimated Capacity: 58.902 G
Space Trigger Util:
Pct Util: 30,9
Pct Migr: 60,0
Pct Logical: 99,6
High Mig Pct: 90
Low Mig Pct: 70
Migration Delay: 0
Migration Continue: Yes
Migration Processes: 1
Reclamation Processes: 1
Next Storage Pool:
Reclaim Storage Pool:
Maximum Size Threshold: No Limit
Access: Read/Write
Description: Primairy Tapepool
Overflow Location:
Cache Migrated Files?:
Collocate?: Group
Reclamation Threshold: 100
Offsite Reclamation Limit:
Maximum Scratch Volumes Allowed: 25
Number of Scratch Volumes Used: 15
Delay Period for Volume Reuse: 0 Day(s)
Migration in Progress?: No
Amount Migrated (MB): 0,00
Elapsed Migration Time (seconds): 0
Reclamation in Progress?: No
Storage Pool Data Format: Native
Copy Storage Pool(s):
Active Data Pool(s):
Continue Copy on Error?: Yes
CRC Data: No
Reclamation Type: Threshold
Overwrite Data when Deleted:
Deduplicate Data?: No
Processes For Identifying Duplicates:
Duplicate Data Not Stored:
Auto-copy Mode: Client
Contains Data Deduplicated by Client?: No


Storage Pool Name: VAULTPOOL
Storage Pool Type: Copy
Device Class Name: LTO3
Estimated Capacity: 29.184 G
Space Trigger Util:
Pct Util: 57,2
Pct Migr:
Pct Logical: 99,6
High Mig Pct:
Low Mig Pct:
Migration Delay:
Migration Continue:
Migration Processes:
Reclamation Processes: 1
Next Storage Pool:
Reclaim Storage Pool:
Maximum Size Threshold:
Access: Read/Write
Description: Vaultpool op
Overflow Location:
Cache Migrated Files?:
Collocate?: No
Reclamation Threshold: 100
Offsite Reclamation Limit: No Limit
Maximum Scratch Volumes Allowed: 50
Number of Scratch Volumes Used: 45
Delay Period for Volume Reuse: 4 Day(s)
Migration in Progress?:
Amount Migrated (MB):
Elapsed Migration Time (seconds):
Reclamation in Progress?: No
Storage Pool Data Format: Native
Copy Storage Pool(s):
Active Data Pool(s):
Continue Copy on Error?:
CRC Data: No
Reclamation Type: Threshold
Overwrite Data when Deleted:
Deduplicate Data?: No
Processes For Identifying Duplicates:
Duplicate Data Not Stored:
Auto-copy Mode:
Contains Data Deduplicated by Client?: No

From the get go we've been having problems as mentioned above. Tapes get mounted in wrong drives for labeling, loading LTO3 scratch for LTO5 stgpool etc.
I noticed the format and mount limit parameters for the legacy LTO devclass were set to DRIVE and DRIVES respectively and figured this might be causing the problems. update it to read ULTRIUM3C and 2 instead but it kept giving us the same problems. We have since moved all the data from the old primary TAPEPOOL_LTO3 to the new TAPEPOOL and deleted old stgpools and devclass but that's not fixed it either.

Would you guys be so kind as to shed some light on this situation? Is what we're trying to achieve at all possible? TS3200 manual says it should be, is TSM up to the job?
Any input is gladly appreciated.

Thanks

[moon-buddy]

Use Logical library setup - one for LTO3 and one for LTO5

The caveat,

You can't have a Domain defined to both. You have to separate your storage pools.

[sjl]

Per http://pic.dhe.ibm.com/infocenter/ts...ons_mixed.html - there are rules and guidelines for mixing drives and generations in LTO libraries. My general advice to clients: if you're mixing things and the maximum difference is one generation (LTO1 and LTO2, or LTO2 and LTO3, or LTO3 and LTO4, or LTO4 and LTO5), you can mix either the drives or the media - not both. So for example, if the library has both LTO4 and LTO5 drives, you should constrain yourself to use only LTO4 media. If the library has only LTO5 drives, you can use both LTO4 and LTO5 media.

The prime point here is that all media should be mountable, readable, and writeable in all drives.

If you're mixing things up by more than one generation's difference (as you are here: LTO3 to LTO5 is two generations), you need to make a complete move to the new technology. Replace all LTO3 drives with LTO5; mark all LTO3 media as read only and check out all LTO3 scratch; and ensure that there is enough LTO5 scratch media to cover day to day operations. Progressively move data off the LTO3 carts and onto LTO5, and check out the LTO3 carts as they're emptied out.

If your budget is such that you can't afford to do that, then moonbuddy's advice is the correct way to go: you need to split the library in two, one with the LTO3 drives and media, the other with the LTO5 drives and media. Trouble is, I don't think the TS3200 has the ability to do that internally - only the TS3310 and TS3500 (aka 3584) do. You might be able to get away with defining the library twice (with the same media changer device), but I'd be very reluctant to try it.

My best suggestion: pull out the LTO3 drives and (optionally) replace them with LTO5, and mark the LTO3 media read only, as I've suggested. There's going to be far too much pain otherwise.

[CoMaboy]

Thanks guys for the info, looks like we will need to switch to LTO5 for the off-site pool as well and eventually get rid of the LTO3 drives and tapes.
I have seen some screenshots from somebody running a similar setup on 6.3.1.0 with and unpartitioned TS3310 and that is running smoothly. His devclasses are set to mountl=drives so i tried that as well. No luck tho, I'm still having the same problem when i need a scratch volume for the LTO5 pool.

Btw, TS3200 does allow partitioning but it is already a small library and I would need to cut it in half.. switching to all LTO5 seems to be the most sensible thing to do. It's the most economical way to get another couple of years life out of our library before we eventually have to move to the TS3310.

[mikeymac]

Thanks guys for the info, looks like we will need to switch to LTO5 for the off-site pool as well and eventually get rid of the LTO3 drives and tapes.
I have seen some screenshots from somebody running a similar setup on 6.3.1.0 with and unpartitioned TS3310 and that is running smoothly. His devclasses are set to mountl=drives so i tried that as well. No luck tho, I'm still having the same problem when i need a scratch volume for the LTO5 pool.

Btw, TS3200 does allow partitioning but it is already a small library and I would need to cut it in half.. switching to all LTO5 seems to be the most sensible thing to do. It's the most economical way to get another couple of years life out of our library before we eventually have to move to the TS3310.

Hi, CoMaboy.

We've had discussions like this many times on this board. I've been running a mixed LTO2/LTO4 drive/media environment for years now with no problems. In my case, we have 3584 tape libraries (only one partition in each physical library). Read up on mixed media/mixed drives in the following Redbook: http://www.redbooks.ibm.com/abstract...6502.html?Open It is the very greatest RedBook that has ever been written.

[CoMaboy]

Hi mikeymac, I have seen you mention that redbook in various other posts while looking for a solution. I think most of it applies regardless of the fact I run TSM on Windows. Still can't get my head around it tho. Who's job is it to detect the format of a scratch volume? TSM or the library?
Not sure if this is normal but the mediatype field in 'select * from libvolumes' is blank for all my tapes, scratch or private. When I check the library inventory it lists the status as Full-Gen. 5 or Full-Gen. 3. Could it be an issue with the library firmware or SCSI driver.

I've opened a ticket with IBM hardware suppport and they couldn't find any problem in the service dump or drive dump I sent them. Opened a software call as well. The guy first said he's had seen this problem before and was sure he cold help me figure it out. Later he sent me an email linking to the document sjl mentioned above. Not supported so case closed.

[sjl]

We've had discussions like this many times on this board. I've been running a mixed LTO2/LTO4 drive/media environment for years now with no problems. In my case, we have 3584 tape libraries (only one partition in each physical library).

My feeling - and it is just a gut feeling - is that the degree of success in this sort of setup (however much I rail against it... ) is directly related to whether or not the library can detect, and pass back, the format of the LTO cartridges. The 3584 definitely can. I don't know if the TS3310 can, but I suspect it can. Similarly, I don't know if the TS3200 can, but I suspect it can't.

TSM should, in principle, be able to figure it out from the tape labels (L3 for LTO3, L5 for LTO5), but it doesn't - and let's face it, tape labels are not reliable indicators anyway. So yeah, that's how it goes. I still stand by my comments with regards to not mixing both, regardless of whether the library "should" be able to handle it - I strongly prefer the simplicity of knowing that every cartridge will mount, read, and write in every drive.

[CoMaboy]

My feeling - and it is just a gut feeling - is that the degree of success in this sort of setup (however much I rail against it... ) is directly related to whether or not the library can detect, and pass back, the format of the LTO cartridges.

When the format does get passed back, where does TSM store this information for scratch volumes?

EDIT:
I found a topic which mentioned a tsm command i didn't know, SHOW SLOTS. It seems TSM has got no help for it. Below is the output I get. Question: what sort of output do you guys get? I get devT=ANY mediaT=-1 for all volumes.

PVR slot information for library TS3200.
Library          : TS3200 
Product Id       : 3573-TL         
Support module   : 2 

Mount count      : 0 

Drives           : 4 
Slots            : 44 
Changers         : 1 
Import/Exports   : 3 

Device           : lb0.1.0.3 

Drive   0, element 256 
Drive   1, element 257 
Drive   2, element 258 
Drive   3, element 259 

Changer 0, element 1 

ImpExp 0, element number 16 
ImpExp 1, element number 17 
ImpExp 2, element number 18 
Slot 0, status Allocated, element number 4096, barcode present, barcode value <CFE778L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 1, status Allocated, element number 4097, barcode present, barcode value <NW4911L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 2, status Allocated, element number 4098, barcode present, barcode value <CFE758L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 3, status Allocated, element number 4099, barcode present, barcode value <NW4913L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 4, status Allocated, element number 4100, barcode present, barcode value <CFE771L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 5, status Allocated, element number 4101, barcode present, barcode value <NW4910L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 6, status Allocated, element number 4102, barcode present, barcode value <CFE769L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 7, status Allocated, element number 4103, barcode present, barcode value <NW4878L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 8, status Allocated, element number 4104, barcode present, barcode value <NW4868L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 9, status Allocated, element number 4105, barcode present, barcode value <CFE774L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 10, status Allocated, element number 4106, barcode present, barcode value <NW4824L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 11, status Allocated, element number 4107, barcode present, barcode value <CFE757L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 12, status Allocated, element number 4108, barcode present, barcode value <CFE772L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 13, status Allocated, element number 4109, barcode present, barcode value <CFE777L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 14, status Allocated, element number 4110, barcode present, barcode value <CFE776L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 15, status Allocated, element number 4111, barcode present, barcode value <CFE775L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 16, status Allocated, element number 4112, barcode present, barcode value <NW4831L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 17, status Allocated, element number 4113, barcode present, barcode value <NW4856L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 18, status Allocated, element number 4114, barcode present, barcode value <CFE764L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 19, status Allocated, element number 4115, barcode present, barcode value <CFE759L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 20, status Allocated, element number 4116, barcode present, barcode value <NW4869L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 21, status Allocated, element number 4117, barcode present, barcode value <NW4863L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 22, status Allocated, element number 4118, barcode present, barcode value <NW4838L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 23, status Allocated, element number 4119, barcode present, barcode value <NW4909L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 24, status Allocated, element number 4120, barcode present, barcode value <CFE761L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 25, status Allocated, element number 4121, barcode present, barcode value <CFE770L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 26, status Allocated, element number 4122, barcode present, barcode value <NW4890L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 27, status Allocated, element number 4123, barcode present, barcode value <NW4884L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 28, status Allocated, element number 4124, barcode present, barcode value <CFE767L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 29, status Allocated, element number 4125, barcode present, barcode value <NW4896L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 30, status Allocated, element number 4126, barcode present, barcode value <NW4891L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 32, status Allocated, element number 4128, barcode present, barcode value <CFE768L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 33, status Allocated, element number 4129, barcode present, barcode value <CFE760L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 34, status Allocated, element number 4130, barcode present, barcode value <NW4902L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 36, status Allocated, element number 4132, barcode present, barcode value <NW4852L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 38, status Allocated, element number 4134, barcode present, barcode value <NW4897L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 39, status Allocated, element number 4135, barcode present, barcode value <NW4916L3>, devT=ANY, mediaT=-1, elemT=ANY
Slot 40, status Allocated, element number 4136, barcode present, barcode value <CFE765L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 41, status Allocated, element number 4137, barcode present, barcode value <CFE766L5>, devT=ANY, mediaT=-1, elemT=ANY
Slot 42, status Allocated, element number 4138, barcode present, barcode value <CFE763L5>, devT=ANY, mediaT=-1, elemT=ANY

slot element range 4096 - 4139

[sjl]

When the format does get passed back, where does TSM store this information for scratch volumes?

I'd be looking at doing a SELECT on the LIBVOLUMES table. I'm fairly sure the detail's in there. Start by doing a SELECT COLNAME FROM COLUMNS WHERE TABNAME='LIBVOLUMES', to see what the columns are; the answer should be reasonably obvious. (Worst case, SELECT * FROM LIBVOLUMES, but that'll probably show you more information than you need; I prefer to trim it down to just the columns of interest.)

SHOW SLOTS is an undocumented command; moderately useful at times, but I'm not sure that that's going to tell you the answer.

You might also get the answer from a QUERY LIBVOLUMES F=D, but don't quote me on that: I'm going from memory, not the manual.

[CoMaboy]

You're right, the select * from libvolumes gives me the exact same information as q libv f=d does. Device Type and Media Type are blank for all my tapes including the scratch tapes. Someone I know using a TS3310 with mixed LTO3 and LTO5 came back to me with the output of SHOW SLOTS on his machine. He gets devT=LTO where I get ANY and mediaT=430 for LTO5 tapes and 417 for LTO3s.

Slot 24, status Allocated, element number 4120, barcode present, barcode value <DAS032L5>, devT=LTO, mediaT=430, elemT=ANY
Slot 25, status Allocated, element number 4121, barcode present, barcode value <DAS163L5>, devT=LTO, mediaT=430, elemT=ANY
Slot 26, status Allocated, element number 4122, barcode present, barcode value <A05332L3>, devT=LTO, mediaT=417, elemT=ANY
Slot 27, status Allocated, element number 4123, barcode present, barcode value <DAS131L5>, devT=LTO, mediaT=430, elemT=ANY
Slot 28, status Allocated, element number 4124, barcode not present, barcode value <>, devT=LTO, mediaT=430, elemT=ANY
Slot 29, status Allocated, element number 4125, barcode present, barcode value <A05383L3>, devT=LTO, mediaT=417, elemT=ANY
Slot 30, status Allocated, element number 4126, barcode present, barcode value <A05395L3>, devT=LTO, mediaT=417, elemT=ANY
Slot 31, status Allocated, element number 4127, barcode present, barcode value <DAS192L5>, devT=LTO, mediaT=430, elemT=ANY
Slot 32, status Allocated, element number 4128, barcode present, barcode value <A05191L3>, devT=LTO, mediaT=417, elemT=ANY

I'm guessing his Q LIBV will show something in the Device Type end Media Type columns. From the looks of it my TSM server is not getting any info on the media apart from the volume name.

[CoMaboy]

By now I have found a workaround for this problem. After checking in LTO5 scratch tapes I manually define them in the LTO5 storage pool. Now I don't have any real LTO5 scratch tapes but rather empty volumes in my primairy storage pool.

Meanwhile I'm still stuck between a rock and a hard place, or rather between a soft and a hardware ticket with IBM. Software support rang today saying they want to close the ticket. TSM doesn't get a proper library inventory from my ts3200 and it does what it can with the info that is available. (TSM seems to receive just the volume label, no device or media type).

A couple of minutes later hardware support called me up as well. They also wanna close the call cause library dumps are clean so from a hardware point of view everything looks ok. He said that the library doesn't communicate any info but the volume label to TSM. It is up to TSM to decide the device and media type using the last 2 digits of the 8 character volume label...

So the answers I get seem rather conflicting and I still don't know which is true. Hw support also suggested I needed to setup TSM with two separate scratch pools, one for every devclass. That was new to me and I don't know if that's possible at all. Mikeymac, would you be able to shed some light on this?

Hw support is now going back to sw support to discuss things further.

[moon-buddy]

As you said, you have a bunch of empty LTO5 tapes which would mean these are of status=private. TSM would not pick this up and use as scratch since it belongs to a storage pool.

So the issue is really Operational in nature.

For TSM to use these empty tapes, these must be of status=scratch.

If you go back to my earlier post, I suggested two logical library setup which is similar to what IBM is suggesting.

[sjl]

Meanwhile I'm still stuck between a rock and a hard place, or rather between a soft and a hardware ticket with IBM. Software support rang today saying they want to close the ticket. TSM doesn't get a proper library inventory from my ts3200 and it does what it can with the info that is available. (TSM seems to receive just the volume label, no device or media type).

Correct.

He said that the library doesn't communicate any info but the volume label to TSM. It is up to TSM to decide the device and media type using the last 2 digits of the 8 character volume label...

Which is not a reliable method of determining the media type. Granted, a well run shop will have "L3" on LTO3 media, and "L5" on LTO5 media, but it's not reliable, and I don't think Tivoli would be particularly interested in putting in a kludge like that - not when a library like the 3584 provides the media information to TSM in a more reliable manner.

So the answers I get seem rather conflicting and I still don't know which is true. Hw support also suggested I needed to setup TSM with two separate scratch pools, one for every devclass. That was new to me and I don't know if that's possible at all. Mikeymac, would you be able to shed some light on this?

No, it's not possible. A scratch pool is assigned to a library, not to a device class. The only way to get two scratch pools is, as moon-buddy says, to split the library into two separate logical libraries.

Ultimately, I think there are only three options.

First: bite the bullet, get rid of the LTO3 drives, and replace them with LTO5. Then checkout all the LTO3 scratch, replace them with LTO5, and gradually move data off your LTO3 tapes and onto LTO5.

Second: just use the LTO5 drives you have (essentially the same as the first option, but with fewer drives - whether this is an issue or not depends on your workload.)

Third: split the library into two logical libraries, and carve up the drives by media type. LTO3 in one logical library, LTO5 in the other. Then define device classes and storage pools according to your needs.

I will be extremely surprised if either the hardware or the software guys come back to you with anything different to the above.