TSM Topics Feed

Wednesday, April 08, 2015

TSM 7.1 Discussion

I was asked by a former colleague to start a discussion on the merits of TSM 7.1. I use TSM 7.1 currently on my newer TSM servers and don't see a huge difference from our 6.4 servers. I know there are some features and updates with 7.1 but none that impact me other than any possible performance enhancements.

So tell me why you upgraded to TSM 7.1 or what features made it a "No Brainer" for you to implement 7.1. I'd like to hear from you.

Q MOUNT Research

So in yesterday's post I spoke of a "better Q MOUNT" using a script I wrote. The problem is the Q MOUNT command does not reference a accessible TSM table. It appears to gathers the information on the mount with the associated process or session from various tables (as seen with Q MOUNT F=D). So I became more intrigued with how TSM gathers the info and decided to bypass TSM all together and research on the DB2 side. Well out of 779 tables in the TSM 7.1 DB2 database I did not see any with MOUNT in the name (you can find a list of all the TSM DB 7.1 tables here). There definietle some tables that appear in the list that are hidden from TSM or unused, but I queried a number of tables looking for any that had fields (columns) that correlated to the Q MOUNT command. The closest I came was the MMS_DRIVE table which has the following fields.

$ db2 "select char(COLNAME,24) as COLNAME,char(TYPENAME,24)as DATATYPE,LENGTH,SCALE from syscat.columns where tabname='MMS_DRIVES' and tabschema='TSMDB1'"

COLNAME                  DATATYPE                 LENGTH      SCALE
------------------------ ------------------------ ----------- ------
ACSDRVID                 VARCHAR                          126      0
CLEANFREQ                INTEGER                            4      0
DEVICE                   VARCHAR                           65      0
DEVTYPE                  INTEGER                            4      0
DEV_INQ                  SMALLINT                           2      0
DRIVENAME                VARCHAR                           31      0
ELEMENT                  INTEGER                            4      0
INQUIRY                  VARCHAR                         1025      0
KBYTES_PROC              BIGINT                             8      0
KBYTES_PROC_HI           INTEGER                            4      0
LIBNAME                  VARCHAR                           31      0
ONLINE                   INTEGER                            4      0
OWNER                    VARCHAR                           65      0
OWNERVOL                 VARCHAR                         1025      0
RD_FORMAT                BIGINT                             8      0
SERIAL                   VARCHAR                           65      0
UPDATE_DATE              TIMESTAMP                         10      6
UPDATOR                  VARCHAR                           65      0
WR_FORMAT                BIGINT                             8      0
WWN                      VARCHAR                           17      0

There should be a table that would show the drive, what tape is mounted, what process or session it's assigned and who owns it, just like a Q MOUNT F=D. I haven't found it yet.

NOTE: Of the 779 tables I listed a more refined search shows 154 tables that correspond to what we seem to be able to access from the TSM admin command line.

Tuesday, March 24, 2015

A Better Q MOUNT?

I was playing around with my QDRV script which I feel gives a better showing than Q DRIVE and realized that I could technically produce a better Q MOUNT display than the built in TSM command provides. FYI - Unlike the Q MOUNT command my QMNT script shows the device definition for the current DRIVE_OWNER, so a storage agent or library client.  I have provided both my scripts for you, let me know what you think. Suggestions are welcome.

QDRV Script

select cast((library_name)as char(15)) as LIBRARY_NAME, -
cast((DRIVE_NAME)as char(16)) as DRIVE_NAME, -
cast((drive_state)as char(10)) as DRIVE_STATE, -
cast((volume_name)as char(8)) as VOL_NAME, cast((online)as char(10)) as ONLINE, -
cast((ALLOCATED_TO)as char(20)) as DRV_OWNER from drives order by library_name, drive_name

QDRV Macro

DEFINE SCRIPT  QDRV DESC="Show tape drive status"
UPDATE SCRIPT  QDRV "select cast((library_name)as char(15)) as LIBRARY_NAME, -"
UPDATE SCRIPT  QDRV "cast((DRIVE_NAME)as char(16)) as DRIVE_NAME, -"
UPDATE SCRIPT  QDRV "cast((drive_state)as char(10)) as DRIVE_STATE, -"
UPDATE SCRIPT  QDRV "cast((volume_name)as char(8)) as VOL_NAME, cast((online)as char(10)) as ONLINE, -"
UPDATE SCRIPT  QDRV "cast((ALLOCATED_TO)as char(20)) as DRV_OWNER from drives order by library_name, drive_name"

QMNT Script

select varchar(a.library_name,15) as LIB_NAME, -
cast((a.DRIVE_NAME)as char(16)) as DRIVE_NAME, -
cast((a.drive_state)as char(10)) as DRIVE_STATE, -
cast((a.volume_name)as char(8)) as VOL_NAME, -
varchar(b.device,12) as device, cast((a.online)as char(10)) as ONLINE, -
cast((a.ALLOCATED_TO)as char(20)) as DRV_OWNER from drives a, paths b -
where a.library_name=b.library_name and a.drive_name=b.destination_name -
and b.source_name=a.ALLOCATED_TO order by a.drive_name

QMNT Macro

DEFINE SCRIPT  QMNT DESC="Show tape mount status"
UPDATE SCRIPT  QMNT "select varchar(a.library_name,15) as LIB_NAME, -"
UPDATE SCRIPT  QMNT "cast((a.DRIVE_NAME)as char(16)) as DRIVE_NAME, -"
UPDATE SCRIPT  QMNT "cast((a.drive_state)as char(10)) as DRIVE_STATE, -"
UPDATE SCRIPT  QMNT "cast((a.volume_name)as char(8)) as VOL_NAME, -"
UPDATE SCRIPT  QMNT "varchar(b.device,12) as device, cast((a.online)as char(10)) as ONLINE, -"
UPDATE SCRIPT  QMNT "cast((a.ALLOCATED_TO)as char(20)) as DRV_OWNER from drives a, paths b -"
UPDATE SCRIPT  QMNT "where a.library_name=b.library_name and a.drive_name=b.destination_name -"
UPDATE SCRIPT  QMNT "and b.source_name=a.ALLOCATED_TO order by a.drive_name"

Wednesday, February 18, 2015

Data Domain Compression

I currently manage 7 Data Domain's (890's and 670's) and none of them are seeing compression above 5x. We obviously need to do some cleanup to get rid of data that is a bad candidate for dedupe, but the question is "where to start?" Have any of you successfully increased you dedupe compression through cleanup? If so what steps did you take?

Tuesday, February 17, 2015

Rumor Mill Update!

So my source at IBM was right except for the name, but she was close, in that IBM is going to be restructuring its storage software group under the Spectrum Storage moniker (Specter was close but Spectrum is a better title). ZDNet states that IBM will be investing $1 Billion in software defined storage which also means TSM is no longer under the Tivoli division....(Please, Please, Please let this be true!) In my opinion Tivoli was a horrible brand name and IBM should have never put ADSM under its umbrella.

Hopefully we will see some huge changes to TSM to make it more competitive. Some cloud connectivity would be a good start as more companies are implementing hybrid data-centers. I'd also like to see the new administration and reporting tool become a little more flexible with user defined reports, but we'll have to wait and see where this restructuring takes TSM or should we now call it Spectrum Storage Manager?

Monday, February 16, 2015

TSM Rumor Mill

So I was talking with an IBM source about IBM's recent layoffs and she stated that many groups were being restructured and TSM was affected. According to my source she said that TSM is possibly going to have a name change. This is no shock for those who remember when TSM was called ADSM under the ADSTAR (ADvanced STorage And Retrieval) division. A partial name drop was something along the lines of Specter blah blah blah!! So take this rumor with some reservations. IBM is going through a major upheaval as they decide what their service path with be in the future, so it's anybodies guess where TSM will end up.