Veritas-bu

[Veritas-bu] Three minutes, then the restore dies (positionin g timeout??)

2003-08-15 11:15:35
Subject: [Veritas-bu] Three minutes, then the restore dies (positionin g timeout??)
From: Mark.Donaldson AT experianems DOT com (Donaldson, Mark)
Date: Fri, 15 Aug 2003 09:15:35 -0600
Yup, I set those equal (actually hard-linked the filenames)... didn't work.


The default values for NUMBER_DATA_BUFFERS & NUMBER_DATA_BUFFERS_RESTORE are
different.  I can't find any documentation that says the *must* be the same
but differing default values implies there's no requirement to match them.

I just successfully restored 1.6 TB just before this with them at different
values.  The mistake might be confusing a buffer count with a buffer size
and/or a block size.  It should perform better with more buffers (to some
plateau), though.

Anyway, I've written this off as a bad spot on my tape.  It won't bpverify,
it won't duplicate, & it won't restore.  The only thing left to try is to
upgrade the drive firmware based on another poster's experiences with
restore success.  Our's turns out to be pretty old.

Thanks for the help.
-M


-----Original Message-----
From: Johan Redelinghuys [mailto:jredelinghuys AT stortech.co DOT za]
Sent: Wednesday, August 13, 2003 1:45 AM
To: Donaldson, Mark; Veritasbu (E-mail)
Subject: RE: [Veritas-bu] Three minutes, then the restore dies
(positioning timeout??)


Hi

Do you use SIZE_DATA_BUFFERS & NUMBER_DATA_BUFFERS? If you do, make sure
that you add (if it is not there), NUMBER_DATA_BUFFERS_RESTORE. This file's
value (eg 32 or 64) must be the same value as NUMBER_DATA_BUFFERS. Try this
and let us know if this helped.

JR

-----Original Message-----
From: Donaldson, Mark [mailto:Mark.Donaldson AT experianems DOT com]
Sent: 12 August 2003 19:59
To: 'Johnny Oestergaard'; Donaldson, Mark; Veritasbu (E-mail)
Subject: RE: [Veritas-bu] Three minutes, then the restore dies
(positioning timeout??)


These are LTO Ultrium drives.  It's the same tape drive each time but a
different drive with each attempt I'm making now.

My next idea is to try to duplicate the tape, using a different media
server, & then try to restore from the duplicate.

The precise three minute apparent (?) timeout on positioning is what makes
this suspicious.  If it were 3:39, then I'd suspect the tape itself more.

-M

-----Original Message-----
From: Johnny Oestergaard [mailto:johnny AT joe DOT net]
Sent: Tuesday, August 12, 2003 11:51 AM
To: Donaldson, Mark; Veritasbu (E-mail)
Subject: Re: [Veritas-bu] Three minutes, then the restore dies
(positioning timeout??)


Is it on the same tape each time?
Is it on the same tapedrive each time?

The only time I have seen media positioning errors in our installations is 
has been due to tape errors.
But sometimes it helped just to load the tape on a different drive.

I have only had the problem on backups (thank God never on restores)

What tapedrives do you use?

/johnny

At 10:56 12-08-2003 -0600, Donaldson, Mark wrote:
>I'm having trouble with a restore.  It seems to be timing out on media
>positioning after exactly three minutes.  Here's the top of the logfile:
>
>   Restore started 08/12/2003 11:38:14
>   11:38:17 (99312.xxx) Restore job id 99312 will require 1 image.
>   11:38:17 (99312.xxx) Media id C00194 is needed for the restore.
>   11:38:20 (99312.001) Restoring from image created Fri Jul 25 13:55:39
2003
>   11:38:22 (99312.001) INF - Waiting for mount of media id CG0194 on
server
>XXX.
>   11:39:04 (99312.001) INF - Waiting for positioning of media id CG0194 on
>server XXX.
>   11:42:04 (99312.001) The following files/folders were not restored:
>   <snip filelist>
>   11:42:09 (99312.001) Status of restore from image created Fri Jul 25
>13:55:39 2003 = media position error
>   11:42:10 (99312.xxx) INF - Status = the restore failed to recover the
>requested files.
>
>If you notice, the time between the positioning start & the failure
>notification is precisely three minutes.  Four attempts, each fails at the
>three-minute mark.
>
>The MEDIA_MOUNT_TIMEOUT, according to the Sysadmin Guide, is supposed to
>control this timing (load plus positioning combined) and I have it set for
>7200 seconds (2 hours) in the master server's bp.conf file.
>
>   > bpconfig -U
>   <snip>
>   Media Mount Timeout:          120 minutes
>   Shared Media Mount Timeout:   120 minutes
>   <snip>
>
>This is v4.5 MP4 on Sol 8.  The client for the restore job is one of four
>media servers (counting the master=media).
>
>Any ideas?
>
>-M
>_______________________________________________
>Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
>http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu