Networker

Re: [Networker] Excessive time to reposition an LTO tape for writting

2003-02-22 13:05:20
Subject: Re: [Networker] Excessive time to reposition an LTO tape for writting
From: Carl Farnsworth <carl.farnsworth AT DIGIDYNE DOT CA>
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Date: Sat, 22 Feb 2003 13:05:15 -0500
Are any of these drives being shared with another storage node?  One
possibility would be that one system is writing to these clone tapes with a
64K block size, for example, and another system trying to use a different
block size, which should result in messages like "positioning by record
disabled" - in which case the forward spacing would eventually get there,
but after a VERY long time.

Are there any pertinent error messages in the daemon.log when the 2nd tape
mount is started - i.e. a mount timeout error?

HTH
Carl Farnsworth
DigiDyne Inc.
519-873-0270

On Fri, 21 Feb 2003 12:20:18 -0500, Jon Hickmott <jonny AT UWO DOT CA> wrote:

>Our Environment
>
>   * Sun E450 running solaris 8 and legato 6.1.1
>   * Fiber attached StorageTek L700 with 4 IBM LTO tape drives
>
>Problem situation
>
>   * running a saveset clone
>   * clone pool has close to full tapes marked as '100%' but not 'Full'
>     (>100GB on then already (full seems to be ~140GB))
>   * The system mounts one of these close to full clone tapes to append
>     to it
>   * It sits there saying "Moving forward xxx files" for 45 minutes or
>     more
>   * In the mean time (about 30 minutes in) the clone process or mount
>     process times out and restarts
>   * A 2nd tape for the clone pool is mounted (if it is close to full
>     the process begins again)
>   * The clone completes on the 2nd mounted tape (or 3rd sometimes)
>   * The other close to full tapes that were mounted end up in a "ready
>     of writing, idle" state after 45 minutes or so and we can not get
>     them back without restarting networker.
>
>Does anyone know why it is taking so long to reposition a tape? Does it
>have something to do with the 100% full indication or the pool
>type(clone)? I have not seen this problem yet while doing regular
>backups - only the clones. Any tips would be great.
>
>Thanks,
>Jon

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=