For backup and archive, RESOURCEUTILIZATION does not map to the number of
drives that will be used. Rather, it provides a guideline for the number
of resources the client should use. Based on the real-time performance
characteristics and the value of RESOURCEUTILIZATION, the client will
increase or decrease the number of resources it uses. The purpose of this
setting is to not simply dictate the number of sessions to use, but to put
a little intelligence behind the number of sessions that will be used.

In particular, the resources are the number of producer sessions (sessions
that figure out what to back up) and the number of consumer sessions
(sessions that actually back up or archive the data). The higher the
RESOURCEUTILIZATION value, the more producer/consumer sessions the client
*may* use. If the system is starved for other resources, or the number of
files to process does not warrant it, then a larger number of sessions may
not be used, even with a large RESOURCEUTILIZATION value.

With a RESOURCEUTILIZATION value of 10, the client will use up to four
producer sessions and four consumer sessions. The idle sessions you saw
were the producer sessions (probably scanning the client file system for
work to do) and the sessions with tapes were the consumer sessions. With
RESOURCEUTILIZATION set to 9, you will get up to three producer sessions
and four consumer sessions. A value of 10 will give you up to four
producer sessions and four consumer sessions.

If you need tighter control over the number of sessions that will be used
(i.e. you *must* have four sessions sending data) , then try dividing the
file systems as evenly as possible into four groups, and creating a script
to launch four separate instances of the client, each instance backing up
a particular set of file systems.

Regards,

Andy

Andy Raibeck
IBM Software Group
Tivoli Storage Manager Client Development
Internal Notes e-mail: Andrew Raibeck/Tucson/IBM@IBMUS
Internet e-mail: storman AT us.eyebm DOT com (change eye to i to reply)

The only dumb question is the one that goes unasked.
The command line is your friend.
"Good enough" is the enemy of excellence.




Kent Monthei <Kent.J.Monthei AT GSK DOT COM>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
11/25/2002 17:29
Please respond to "ADSM: Dist Stor Manager"


        To:     ADSM-L AT VM.MARIST DOT EDU
        cc:
        Subject:        ResourceUtilization in TSM 5.x client - have you tried 
it,         has it
improved performance, do you have concerns/problems using it ?



I've been trying to make use of ResourceUtilization to backup a 1TB Oracle
database directly to tape (4 drives), but am just about ready to give up
on it.

My objective is to have the TSM Client open 4 concurrent data sessions,
each backing up a separate filespace going to a separate tape drive,
utilizing all 4 tape drives 100% of the time.  My plan was to bind all 20
filespaces to an MC whose CopyGroup is directed to a tapepool, and set
client ResourceUtilization to 5.  I expected the TSM Client to initiate 4
concurrent data sessions/threads, and the TSM Server to mount/use 4 tapes,
1 per thread.

Not so.   With ResourceUtilization=5, TSM Client opened 5 server sessions,
but only 2 sessions went into MediaW status and only 2 tape-mounts were
performed by the server.  2 sessions remained in IdleW status.  I
confirmed all drives/paths were online.

With ResourceUtilization=7, TSM Client opened 7 sessions with the server.
4 went into MediaW state immediately.  However, the server only mounted 1
tape.  3 sessions sat in MediaW and 2 in IdleW status for almost 50
minutes before I terminated the test.  Throughout the test 'q mount'
reported only 1 in-use tape and no additional pending mounts.  There were
no mount-related errors in the TSM Server Activity Log and no sign of
problems in the Client schedlog.

With ResourceUtilization=10 (max), TSM Client opened 8 sessions with the
server.  4 went into MediaW immediately and shortly after that 4 tapes
were mounted and in-use.  This is almost the desired behavior.  However, 3
other sessions sat in IdleW the whole time, and the TSM Client & Server
logs provided very little insight or progress-reporting.   It also
appeared that only one thread/session (the 1st filespace backed up) was
actively appending to the schedlog, even though 4 tapes were
mounted/in-use and Collocation=Filespace was set on the TSM Server
tapepool.

With ResourceUtilization, it seems that I can neither predict nor see what
is actually happening.  On a 1 TB database backup comprised of >20
filespaces that will take up to 24 hours to complete, I absolutely must
have the ability to accurately control and monitor the progress/status of
the backup.  ResourceUtilization isn't giving me that control or insight.

Comments?........rsvp, thanks

Kent Monthei
GlaxoSmithKline