Currently preparing changes to our tape hardware in the Data Center. Current TSM Server is Version 5, Release 1, Level 7.0 on OS390 2.10 MVS Mainframe. We are moving to ZOS sometime before January. The change at this time is our tape environment. Will probably give too much detail here but I need to completely understand the implications of the changes. If anyone can think of something I may have missed in my implementation plan I sure would appreciate a response.
Currently we have two ATL'S,
1. A Sutmyn 5600 Memorex ATL, has 3490 drives and uses 3490 cartridges with 800 MB's
(Currently 98% of this ATL is used just for TSM Archives)
2. An IBM Magstar 3494, has 3590B drives and uses 3590J cartridges (about 12 GB each) all tape racks are full inside this ATL.
3. TSM backup nightly incremental process,
a. The nightly incremental backups go to a DASD primary diskpool.
b. The data is copied to an Offsite copypool (Magstar 3590J) to go offsite.
c. Before the next nightly incremental backup the DASD primary diskpool data is migrated to a primary sequential stgpool on the Magstar (3590J) carts.
d. Backupset's are sent directly to the Magstar (3590J), ejected but kept onsite.
e. All the weekly, monthly, and yearly Archives go directly to tape on the Memorex (3490 carts) and are kept onsite.
The following are the new hardware changes that will have an effect on our TSM archive environment,
1. The Sutmyn 5600 Memorex ATL is getting replaced with an IBM Virtual Tape Server 3494-B18 VTS with 432 GB of cache. It will have four 3590E control units each with 16 drives, for a total of 64. It will connect to an IBM 3494 D12 with four 3590E drives for the VTS. The virtual tape volumes in cache will be defined exactly as the Memorex tape volumes, 3490 cartridge with 800 MB's. All data currently going to the Memorex that does not go offsite will be diverted to VTS cache, this includes all the TSM Archives. Eventually, the data in the VTL cache will migrate to the 3494 D12. They will be 3590E drives and using 3590K cartridges, which should hold between 40-50 GB's each.
a. Supposedly, TSM will not know the difference between the Memorex cartridge volumes and the VTL volumes because as they will be defined the same. As a result of this, my plan for the TSM archives configuration is as follows.
i. Create a new primary sequential storage pool example- ARCHCART02
ii. Keep collocation ON to reduce the number of VTS volumes required for a full retrieve if it was ever requested. Collocation with the VTS will not minimize the physical tapes used but will minimize the logical volumes used.
iii. The TSM archives will be diverted to the new ARCHCART02 stgpool through an MVS system change, TSM should never know the difference. After archives start going to ARCHCART02, start moving the data from the old ARCHCART to ARCHCART02.
iv. Currently reclamation for archives is done once a week, I should only need to do this once a month during non-peak hours now.
v. I can set a much higher MOUNTLIMIT with defining the device class as there will now be 64 virtual drives available. MOUNTRETENTION can now be set to zero.
I do have a question here about backing up the TSM database. Will it be possible to do some kind of DB Backup to the VTS and still keep my DB Series that is going offsite?
2. In addition to the VTS and our current Magstar, we are adding a 3494 -D14 and a 3494 -D12, that have a total of (6) 3590E drives. As with the Magstar, it will use the 3590J cartridges but will they should hold more data because of the difference between the Magstar 3590B drives vs. the 3590E drives on the new ATL. The following is the how this hardware change will have an effect on our TSM backup environment,
a. The nightly incremental backups will still go to DASD.
b. The OFFSITE sequential copy pool will stay in the current Magstar along with any other data cartridges that need to go offsite or be ejected for some reason.
c. The TSM data on DASD will now migrate to the new ATL with the 3590 E drives. Because the current Magstar tape racks are completely full, we have to eject between 100-200 carts each week make it through the lights-out weekend. Reclamation during the weekdays of both the primary BACKCART01 and OFFSITE copy pool call most of those tapes to be put back in.
d. Although the cartridge sizes are the same, the tape drives are different. My plan for the TSM incremental backup configuration is as follows.
i. Define a new sequential primary storage pool for the new ATL example-BACKCART02
ii. Turn collocation on for faster restores.
iii. The TSM incremental backups will then be migrated to the new BACKCART02 stgpool. After the backup data starts going to BACKCART02, I will set reclamation on the old BACKCART01 to point to BACKCART02.
iv. Because the cartridges in BACKCART02 will hold more than those already created in BACKCART01, we will not put these carts into the new ATL. They can be read but not written to until the data is expired and the tape is scratched. I will move the data from BACKCART01 to BACKCART02 a little bit everyday until done.
Can anyone see a problem with these plans that hasn't occurred to me? Anything I should take into consideration or improve upon? I am open to all ideas & suggestions. I'm hoping to get this all finished before we move to ZOS, at which time we will upgrade our TSM server version to 5.2.
Thank you for your time,
Shannon
Operations Analyst
Madison Gas & Electric Co.
Data Center Services
Information Management Services
- Re: Changes to Tape Management hardware in Data Center Shannon Bach
- Re: Changes to Tape Management hardware in Data Center Steve Harris
- Re: Changes to Tape Management hardware in Data Center Gee, Norman
- Re: Changes to Tape Management hardware in Data Center Shannon Bach
- Re: Changes to Tape Management hardware in Data Center Gee, Norman
- Re: Changes to Tape Management hardware in Data Center Shannon Bach