>Date:    Mon, 5 Mar 2001 08:45:40 -0500
>From:    Richard Sims <rbs AT BU DOT EDU>
>Subject: Re: Longlasting tape-reclamation run
>
>Geoff - Greetings from Up Over.  ;-)
>

Hello Richard,

Hope you are surviving the blizzard - we don't get many of those in Sydney!

>>2. Influence of client type.
>>
>>   I have clients of the following types: Novell Netware, Unix, NT, and
>>also NT with the Lotus Notes agent.  Since I have collocation on my
>>onsite tape pool, I was able to determine that the tapes causing trouble
>>all belonged to Notes clients.  Looking at a list of my tape pool today
>>(about 200 volumes), I can say that for the non-Notes clients, the
>>number of clusters is always less than 10.  The Notes client volumes have
>>HUNDREDS (highest today is 967).
>>   I don't know if this is something to do with the Notes agent itself,
>>or just a result of the fact that Notes seems to generate vast numbers of
>>very small documents.
>
>Though we may have collocation activated in the server, I believe it to be
>the general case that API-based clients either cannot or do not collocate.
>(This is the case with HSM, at least.)  API-based clients which back up
>numerous small client files thus pose a special burden on the server.
>

API clients and collocation:  I'm not quite sure what you mean.  In my
case with the Notes agent, collocation clearly works at the tape end.
I've always assumed that the agent uses the ADSM API to fake up a file for
each Notes document, this faked file then gets sent to the server who will
treat it (for collocation and other purposes) just like any other client
file?

>>   I was able to determine that the process spends most of its time
>>waiting for data from the input tape.  In S/390 terms, every small burst
>>of data transfer (few milliseconds) is followed by a Locate Block (tape
>>search) function which takes more like 10 SECONDS.
>
>What tape technology was involved?  Just curious, as some types are
>bow-wow's in the Start-Stop department.
>

Tape technology: I have STK 9840's which are advertised as being pretty
good
in the load/search/rewind area.  The Locate Block instruction is an S/390
operation which came in with square tapes, so it is possible that in the
SCSI world things happen differently and so maybe ADSM uses code which
doesn't
port very well to the S/390? (just guessing, noone from Tivoli/IBM/STK said
anything like this to me).

>>5. How do I live with it?
>
>I guess the ultimate solution is to not do Reclamations.  We should all
>have libraries with so many tapes that the files just age out so that
>the tapes go empty and return to scratch by themselves.  And we should
>all be getting six-figure salaries - to the left of the decimal point.

Doing without reclamation? Management here already have a perception that
9840 cartridges are expensive, so I would hesitate to apply for an infinite
supply. (might be easier to get the six figures!).

With "incremental forever", "forever" can be a long time.


Many thanks for your interest, and of course for your continuing
contributions
to the list.

Cheers,
Geoff