Betr: Re: FW: Why are tape mounts being requested?

     Ok,

     I don't no if anyone has mentioned this yet, but have you checked the
     maximum file size for the disk pool, if clients are sending files
     exceeding the max file size then they're routed thru to the next
     storage pool.

     Regards, groeten,

     Walter Ridderhof

     Mainland Sequoia
____________________________ Antwoord-separator ________________________________
Onderwerp: Re: FW: Why are tape mounts being requested?
Auteur:  "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU> in inet-1
Datum:    11/24/98 8:29 AM


Hmmm...

Sounds like there is not enough space in your disk pool.  What might
be happening

the disk pool is filling up
it reaches the high water mark and migration starts
the migration process empties out the disk pool to tape
the incoming data is coming in faster than it can be copied out
the disk storage pool fills.

At that point, all the clients that were going to the disk pool want
to go to tape, courtesy of the "Next Storage Pool" setting.

All the clients block waiting for a tape mount and do not move back to
the disk pool even if that has enough space.

The effect will be what you saw: 25 clients sitting there waiting for
tapes. <sigh> They will only complete one at a time as a tape drive
becomes free.


Workarounds?

1.  Are the tape mounts happening too slow?  Get rollerskates for your
ops ;)

2.  Try using lower migration thresholds.  This will start off the
migrations earlier and let them run for longer.  This may drain the
disk pool faster.

3.  More disk space for your staging pool

4.  Spread the client schedules out if you can.  Make the data arrive
slower.

5.  If you have a few really large file arriving, you could set the
"Maximum Size Threshold" or a management class to direct them to tape.


Russell



[Charset iso-8859-1 unsupported, filtering to ASCII...]
> I think you mis-understood my note -- I did the migration threshold
> at 0 during the afternoon to clean everything out, and it was set
> back to hi=95 lo=80 at 18:45, 15 minutes before the nightly
> backups kick in. And none of the clients had problems needing
> tapes until 2 hours later.
>
> Linnea Nichols
>
> Fairfax Co. Govt, Fairfax, VA    Linnea.Nichols AT co.fairfax.va DOT us
> Phone:703/324-2708                FAX: 703/324-3931
>
> -----Original Message-----
> From: Kelly J. Lipp [mailto:lipp AT STORSOL DOT COM]
> Sent: Monday, November 23, 1998 11:22 AM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Re: Why are tape mounts being requested?
>
>
> Having the threshold at 0 during client backups is not a good idea (as
> you've learned).  You want to perform the migration during the day, while
> there are no client backups running.  The worst of this, is setting the
> threshold backup up while the client backups are waiting for tapes won't
> change anything.  You're stuck until a) the backups complete, or b) you
> kill them all with cancel session.
>
> Kelly Lipp
> Storage Solutions Specialists, Inc.
> lipp AT storsol DOT com
> www.storsol.com
> (719) 531-5926
>
> -----Original Message-----
> From:   Nichols, Linnea [SMTP:linnea.nichols AT CO.FAIRFAX.VA DOT US]
> Sent:   Monday, November 23, 1998 8:43 AM
> To:     ADSM-L AT VM.MARIST DOT EDU
> Subject:        Re: Why are tape mounts being requested?
>
> Things are looking worse than I realized. I checked my copygroups, and
> all the destinations point correctly to the BACKUPPOOL (which is my
> disk pool). On Friday afternoon, I changed the thresholds on this pool to
> hi=0 lo=0 and migrated everything off. I then did a MOVE DATA on every
> single one of the files making up that pool. I also added a Q SESSION
> display every hour -- (and yes, I have caching turned on for this pool).
> BACKUPPOOL is set with Maximum Size Threshhold No Limit.
>
> At 18:45, I do a  "UPDATE STGPOOL backuppool hi=95 lo=80 migpr=1"
>
> Standard nightly backups start at 19:00. About 2 hours into the nightly
> backups, I see the first tape mount kick in. Q SESSION shows 6
> sessions backing up, and only 1 needing a  tape. An hour later,
> everything is a bogged mess. I have 27 sessions in progress, 2 of
> them using disk, and 25 of them waiting for/using tapes!!!!! At 00:10, a
> migration kicks off.
>
> Tonite, I'll add a "Q STGPOOL BACKUPPOOL f=d" to my diagnostics
> to see if I can figure anything else out -- In the meantime, my operators
> are very upset with all the tape mounts that are occuring, and it's
> impacting
> out regular MVS batch production as we have a shortage of drives right now.
>
> Any more ideas?? (many thanks for the ones so far) I keep hoping it's
> something dumb on my part, but I'm beginning to think it may be a bug.
>
>
>
> Linnea Nichols
>
> Fairfax Co. Govt, Fairfax, VA    Linnea.Nichols AT co.fairfax.va DOT us
> Phone:703/324-2708                FAX: 703/324-3931