This is a multi-part message in MIME format.
------=_NextPart_000_000B_01C2B807.E5DB1FA0
Content-Type: text/plain;
charset="iso-8859-7"
Content-Transfer-Encoding: quoted-printable
HI,
I think that this is netbackup problem not atl problem.
ATL tries to move the tape from the drive but the tape isn't unloaded=20
(if you try to reset the drive from device monitor you will fail and the =
drive will go down)=20
The AIX command to unload the drive works fine.
Our solution is to use the AIX command to unload all drives after a =
backup (bpend_notify) using the AIX command . The drive will unload the =
tape ,then ATL will find the tape unloaded and will move it to the right =
bin. This works for us (all other drives working with backups will not =
disterb from this command (use scsi_reserved).
Netbackup 4.5 has a drive_unload_notify in volmgr/database. )ne of the =
parameters is the drive being unload. it is beter to use this script to =
unload with AIX command the spesific drive and not all .(this is not =
tested. the customer is in 4.3 )
Monovasios Stefanos
Performance Technologies
Greece
Subject: RE: [Veritas-bu] AIX/Quantum/Fibre
Date: Thu, 9 Jan 2003 12:12:23 +0200
From: "Daniel Bass" <daniel AT mbi.co DOT il>
To: "David A. Chapa" <david AT datastaff DOT com>,
<veritas-bu AT mailman.eng.auburn DOT edu>
David
We have had similar problem with one of our customers with very similar
environment (ATL p7000 library with SDLT drives attached to AIX server
via fiber bridge and switch.)
Following ATL suggestion we changed "AUTO_UNLOAD_ENABLE" parameter to
"yes" and it helped us to avoid unload problems.
This ATL library parameter force robot to perform unload command every
time the robot is performing move cartridge from the drive.
Check with ATL if this parameter is applicable in your case.
-----Original Message-----
From: David A. Chapa [mailto:david AT datastaff DOT com]=20
Sent: Wednesday, January 08, 2003 9:49 PM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] AIX/Quantum/Fibre
Importance: High
Customer has an F80 AIX system running 4.3.3 with all patches.
This system
has 2 IBM 6228 fiber cards in it and they are attached to a McData 24
port
switch and then they have a P3000 8 drive DLT robotic unit attached to
the
switch via fiber. Their issue is that 50% of the time we can
mount/unmount
tapes from the drives with no issues and backups run fine BUT the other
50%
of time the tape does not get unmounted and therefore the move command
cannot find the tape in the drive and the drive goes down. IBM stated
that
this is an Quantum P3000 fiber brigde issue and not their problem.
Quantum
has stated that the fiber cards we are utilizing are the issue. Quantum
is
working on this issue via firmware upgrades but have not produced a
solution
yet.
The above was the info I receive from my customer.
Additionally, IBM told them that the OS has a pseek routine that queries
the
drives to see if they are available. Quantum's Fibre bridges has this
already built-in, so apparently the thoughts are that the repeated pseek
is
overloading the fibre bridge causing it to reset.
Any one see this before? Is there a temporary WORKAROUND?
Thanks
David
------=_NextPart_000_000B_01C2B807.E5DB1FA0
Content-Type: text/html;
charset="iso-8859-7"
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML><HEAD>
<META http-equiv=3DContent-Type content=3D"text/html; =
charset=3Diso-8859-7">
<META content=3D"MSHTML 6.00.2600.0" name=3DGENERATOR>
<STYLE></STYLE>
</HEAD>
<BODY bgColor=3D#ffffff>
<DIV><FONT face=3DArial size=3D2>HI,</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2>I think that this is netbackup problem =
not atl=20
problem.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2>ATL tries to move the tape from =
the drive but=20
the tape isn't unloaded </FONT></DIV>
<DIV><FONT face=3DArial size=3D2>(if you try to reset the drive from =
device monitor=20
you will fail and the drive will go down) </FONT></DIV>
<DIV><FONT face=3DArial size=3D2>The AIX command to unload the =
drive works=20
fine.</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2>Our solution is to use =
the AIX command to=20
unload all drives after a backup (bpend_notify) using the AIX =
command=20
. The drive will unload the tape ,then ATL will find the tape unloaded =
and will=20
move it to the right bin. This works for us (all other drives =
working with=20
backups will not disterb from this command (use =
scsi_reserved).</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2>Netbackup 4.5 has a drive_unload_notify =
in=20
volmgr/database. )ne of the parameters is the drive being unload. it is =
beter to=20
use this script to unload with AIX command the spesific drive and not =
all .(this=20
is not tested. the customer is in 4.3 )</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2>Monovasios Stefanos</FONT></DIV>
<DIV><FONT face=3DArial size=3D2>Performance Technologies</FONT></DIV>
<DIV><FONT face=3DArial size=3D2>Greece</FONT></DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV><FONT face=3DArial size=3D2></FONT> </DIV>
<DIV>Subject: RE: [Veritas-bu] AIX/Quantum/Fibre<BR>Date: Thu, 9 Jan =
2003=20
12:12:23 +0200<BR>From: "Daniel Bass" <<A=20
href=3D"mailto:daniel AT mbi.co DOT il">daniel AT mbi.co DOT il</A>><BR>To:
"David =
A. Chapa"=20
<<A=20
href=3D"mailto:david AT datastaff DOT com">david AT datastaff DOT
com</A>>,<BR> =
; =20
<<A=20
href=3D"mailto:veritas-bu AT mailman.eng.auburn DOT edu">veritas-bu AT mailman
DOT eng.=
auburn.edu</A>><BR><BR>David<BR>We=20
have had similar problem with one of our customers with very=20
similar<BR>environment (ATL p7000 library with SDLT drives attached to =
AIX=20
server<BR>via fiber bridge and switch.)<BR>Following ATL suggestion we =
changed=20
"AUTO_UNLOAD_ENABLE" parameter to<BR>"yes" and it helped us to avoid =
unload=20
problems.<BR>This ATL library parameter force robot to perform unload =
command=20
every<BR>time the robot is performing move cartridge from the =
drive.<BR>Check=20
with ATL if this parameter is applicable in your =
case.<BR><BR>-----Original=20
Message-----<BR>From: David A. Chapa [mailto:david AT datastaff DOT com] =
<BR>Sent:=20
Wednesday, January 08, 2003 9:49 PM<BR>To: <A=20
href=3D"mailto:veritas-bu AT mailman.eng.auburn DOT edu">veritas-bu AT mailman
DOT eng.=
auburn.edu</A><BR>Subject:=20
[Veritas-bu] AIX/Quantum/Fibre<BR>Importance: High<BR><BR><BR>Customer =
has an=20
F80 AIX system running 4.3.3 with all patches.<BR>This system<BR>has 2 =
IBM 6228=20
fiber cards in it and they are attached to a McData 24<BR>port<BR>switch =
and=20
then they have a P3000 8 drive DLT robotic unit attached =
to<BR>the<BR>switch via=20
fiber. Their issue is that 50% of the time we=20
can<BR>mount/unmount<BR>tapes from the drives with no issues and backups =
run=20
fine BUT the other<BR>50%<BR>of time the tape does not get unmounted and =
therefore the move command<BR>cannot find the tape in the drive and the =
drive=20
goes down. IBM stated<BR>that<BR>this is an Quantum P3000 fiber =
brigde=20
issue and not their problem.<BR>Quantum<BR>has stated that the fiber =
cards we=20
are utilizing are the issue. Quantum<BR>is<BR>working on this =
issue via=20
firmware upgrades but have not produced a<BR>solution<BR>yet.<BR><BR>The =
above=20
was the info I receive from my customer.<BR><BR>Additionally, IBM told =
them that=20
the OS has a pseek routine that queries<BR>the<BR>drives to see if they =
are=20
available. Quantum's Fibre bridges has this<BR>already built-in, =
so=20
apparently the thoughts are that the repeated pseek<BR>is<BR>overloading =
the=20
fibre bridge causing it to reset.<BR><BR>Any one see this before? =
Is there=20
a temporary WORKAROUND?<BR><BR>Thanks<BR><BR>David</DIV></BODY></HTML>
------=_NextPart_000_000B_01C2B807.E5DB1FA0--
|
