Giles beat me to it. I tested maxtapebuf with combined store/slt back in
2005, and it helped a lot.
I have plain-old turbostore, without the "7x24 True-Online Backup", so I
don't think the compress option applies to us. But I think I have the tape
drive (dds3) set up to do hardware compression with "DEVCTRL.MPEXL.TELESUP 7
ENABLE DISABLE".
Except for the large number of small files, I think we would qualify as a
medium-small site. We have an A500 with just the two internal drives and
discfree says they are 78% available. Nightly full backup fits on one dds3.
----- Original Message -----
From: "Craig Lalley" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Monday, May 18, 2009 15:13
Subject: Re: [HP3000-L] backup hang and devinfo hang
I always thought the SLT tape with store was not very efficient, I don't
think the compress, maxtapebuf parameters were supported?
-Craig
--- On Mon, 5/18/09, donna hofmeister <[log in to unmask]> wrote:
From: donna hofmeister <[log in to unmask]>
Subject: Re: backup hang and devinfo hang
To: [log in to unmask]
Date: Monday, May 18, 2009, 4:55 PM
Imo -- combining a (full) backup onto an slt tape is only viable on 'small'
systems (think 918).
On 'medium' size systems, you might be able to do something like a
store-to-disk (using whatever archiving product your site uses) as one
process. Then as a separate process (job), include the std file on your
slt. *if* you are using a 3rd-party archiving too then you'll want to take
care to make sure you include that software on the slt as well!
On 'large' systems -- make a nice simple slt (or minimal) and get your
backup out to tape/dlt.
- d
---
Donna Hofmeister
Allegro Consultants, Inc.
408-252-2330
> -----Original Message-----
> From: HP-3000 Systems Discussion [mailto:[log in to unmask]] On
> Behalf Of Dave Powell, MMfab
> Sent: Monday, May 18, 2009 2:36 PM
> To: [log in to unmask]
> Subject: Re: [HP3000-L] backup hang and devinfo hang
>
> :file t = $null
> :store mysmallfileset; *t;
> STORE/RESTORE VERSION C.75.09 (C) 1986 HEWLETT-PACKARD CO.
>
>
>
> ----- Original Message -----
> From: "Craig Lalley" <[log in to unmask]>
> To: <[log in to unmask]>
> Sent: Monday, May 18, 2009 14:00
> Subject: Re: [HP3000-L] backup hang and devinfo hang
>
>
> Just wondering, did you post your version of STORE?
>
> To find it, just store a file (to disk) and look at the banner.
>
> -Craig
>
>
> --- On Mon, 5/18/09, Dave Powell, MMfab <[log in to unmask]> wrote:
>
> From: Dave Powell, MMfab <[log in to unmask]>
> Subject: Re: backup hang and devinfo hang
> To: [log in to unmask]
> Date: Monday, May 18, 2009, 3:31 PM
>
> Ouch !
>
> Presumably it is only a very intermittent error, since we currently
> backup
> 80,000+ files a night on a combined store/slt, and we have only had
> this
> error once. Also, in our case we did not get the "Program terminated in
> an
> error state. (CIERR 976)" -- both the backup job and devinfo in the
> other
> job hung. Actually, the "hang" part bothered us a lot more than an
> abort
> would have.
>
> We had good old fashioned tape errors a few days before, and again a
> few
> days later, but nothing since the operator gave the drive a good
> cleaning,
> so I was *guessing* a cause along those lines.
>
> FWIW, the number of files in our backups increases steadily, mostly
> because
> we archive several hundred reports a month. 82,000 Friday night, 65,000
> in
> 09/20078, 37000 in 11/2003, etc.
>
> So, is there some point at which adding approx 800+ files a month gets
> us
> into REAL trouble ??
> (they're in separate groups, so a limit on files in a group should
> never be
> an issue).
>
>
> Dave (RELEASE: C.75.00 MPE/iX HP31900 C.45.05 USER VERSION: C.75.05)
> Powell,
> MMfab
>
>
> ----- Original Message ----- From: "cathlene mcrae"
> <[log in to unmask]>
> To: "HP3000 List" <[log in to unmask]>; "Dave Powell"
> <[log in to unmask]>
> Sent: Monday, May 18, 2009 09:17
> Subject: Re: backup hang and devinfo hang
>
>
> Problem is a know issue, with sysgen and store. If the number of files
> for
> the
> store is > 32k files.
>
> Cathlene Mc Rae
> HP SR. Response Center Engineer
> [log in to unmask]
>
>
>
> When creating a combined SLT/STORE tape, the Store process aborts with:
>
>
> SENDMAIL TO SYSGEN FROM STORE FAILED. (S/R 1741)
> **error** failed to receive status from STORE--
> Can't make tape.
> **error** Tape command failed.
> SYSGEN tape is NO GOOD!
> sysgen> e
>
>
>
>
> Program terminated in an error state. (CIERR 976)
> REMAINDER OF JOB FLUSHED.
>
> SENDMAIL TO SYSGEN FROM STORE FAILED. (S/R 1741)
>
> This causes SYSGEN to mark the tape as bad and unreadable.
>
> SOLUTION:
> >From original document 4701321604 (originally seen on 5.5):
>
> Cause:
>
> One of the fields in the sendmail record that is passed between SYSGEN
> and
> STORE is a 16-bit field that is used to contain the number of files
> that
> were
> successfully stored. If the number of files that are successfully
> stored is
> greater than 32767, the process will fail and the tape will be marked
> as bad
> .
>
> Temporary Solution:
>
> The only temporary workaround is to store less than 32K files when
> creating
> a
> combined SLT/STORE tape.
>
> * To join/leave the list, search archives, change list settings, *
> * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
>
> * To join/leave the list, search archives, change list settings, *
> * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
>
> * To join/leave the list, search archives, change list settings, *
> * etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|
