Networker

Re: [Networker] Tape block sizes (?) on Linux, Solaris, and NetApp

2004-08-13 10:52:58
Subject: Re: [Networker] Tape block sizes (?) on Linux, Solaris, and NetApp
From: "Reed, Irene" <Irene.Reed AT TEA.STATE.TX DOT US>
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Date: Fri, 13 Aug 2004 09:52:43 -0500
Have you recently upgraded to 7.1.1 or 7.1.2 of Legato on the Sun or
Linux box?  I only ask as I have a Netapp and Sun that use DDS and once
I upgraded the Sun box to 7.1.1 and now at 7.1.2 I have had numerous
tape issues.

Irene Reed

-----Original Message-----
From: Legato NetWorker discussion [mailto:NETWORKER AT LISTMAIL.TEMPLE DOT EDU]
On Behalf Of Rich Graves
Sent: Friday, August 13, 2004 9:49 AM
To: NETWORKER AT LISTMAIL.TEMPLE DOT EDU
Subject: [Networker] Tape block sizes (?) on Linux, Solaris, and NetApp

We have a backup SAN with 2 NetApps, 2 Suns, and 1 Linux dedicated
storage
node (licensed to back up only itself) sharing 7 AIT-3 drives, 3 of
which
are enabled for Dynamic Drive Sharing. The Suns are configured

  SONY_AIT   =   1,0x36,0,0xd639,4,0x00,0x00,0x00,0x00,0;

which I think means that they will use full 64K block size (Sun's
default
behavior is 64K - 1).

The NetApps are at the defaults.

The NetApps and Suns have interoperated happily for years.

I'm not sure we've pinned it down, but I think the Suns and NetApps are
having intermittent problems reading tapes originally labeled on the
Linux
box, and vice versa.

On the Linux box, dmesg complains like

st0: Failed to read 61440 byte block with 32768 byte read.
st0: Failed to read 61440 byte block with 12288 byte read.
st0: Failed to read 61440 byte block with 32768 byte read.
st0: Failed to read 61440 byte block with 32768 byte read.
st0: Failed to read 61440 byte block with 12288 byte read.
st0: Failed to read 61440 byte block with 32768 byte read.
st1: Failed to read 61440 byte block with 32768 byte read.
st1: Failed to read 61440 byte block with 12288 byte read.
st1: Failed to read 61440 byte block with 32768 byte read.
st1: Failed to read 61440 byte block with 32768 byte read.

I don't understand why anything is trying to do a 12288 read at all?

What should I be trying to troubleshoot?

The ultimate errors on the Solaris side are

08/13/04 06:24:07 nsrd: media warning: /dev/rmt/16ubn reading: Not
enough space
08/13/04 06:24:23 nsrd: media warning: /dev/rmt/16ubn reading: Tape
label read: Not enough space
08/13/04 06:57:34 nsrd: media warning: /dev/rmt/14ubn reading: Not
enough space
08/13/04 06:57:50 nsrd: media warning: /dev/rmt/14ubn reading: Tape
label read: Not enough space

The Linux box does not log anything in daemon.log other than healthy
nsrmmd
starts and stops.

The daemon.log errors relating to the NetApps are:

08/13/04 03:02:00 nsrd: media info: loading volume A00621 into
rd=dop:nrst18a
08/13/04 03:02:27 nsrd: rd=dop:nrst18a Verify label operation in
progress
08/13/04 03:02:34 nsrmmd #14: ndmp tape bsf failed
08/13/04 03:02:48 nsrd: media warning: rd=dop:nrst18a reading: no tape
label found
08/13/04 03:02:49 nsrd: rd=dop:nrst18a Eject operation in progress
08/13/04 03:03:39 nsrd: Jukebox 'Qualstar' failed: expected volume
'A00621' got 'NULL'.
08/13/04 03:04:28 nsrd: media info: suggest mounting A00400 on dop for
writing  to pool 'NDMP'
08/13/04 03:04:29 nsrd: media info: loading volume A00400 into
rd=dop:nrst18a
08/13/04 03:04:47 nsrd: rd=dop:nrst18a Verify label operation in
progress
08/13/04 03:04:54 nsrmmd #14: ndmp tape bsf failed
08/13/04 03:05:55 nsrd: media warning: rd=dop:nrst18a reading: no tape
label found
08/13/04 03:05:56 nsrd: rd=dop:nrst18a Eject operation in progress

On Linux, the tape says:

[root@moe root]# tapeinfo -f /dev/sg20
Product Type: Tape Drive
Vendor ID: 'SONY    '
Product ID: 'SDX-700C        '
Revision: '0103'
Attached Changer: No
SerialNumber: '0001326788'
MinBlock:2
MaxBlock:16777215
SCSI ID: 0
SCSI LUN: 1
Ready: no

What more should I be looking at to troubleshoot/fix this?
--
Rich Graves <rcgraves AT brandeis DOT edu>
UNet Systems Administrator

--
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.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=

--
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.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=