This is a multipart message in MIME format.
--=_alternative 0070BA4986256F20_=
Content-Type: text/plain; charset="us-ascii"
----- Forwarded by Michael F Lavelle/LAKE/CHMS/ABBOTT on 10/01/2004 03:30
PM -----
Michael F Lavelle
10/01/2004 11:47 AM
To: veritas-bu-admin AT mailman.eng.auburn DOT edu
cc: Michael F Lavelle/LAKE/CHMS/ABBOTT@ABBOTT
Subject: Volume Expiration and ScratchPool
Hi,
I don't want to start a shoot-from-the-hip philosophical
discussion, but my topic is Volume Expiration.
I understand the treatment of tape expiration varies widely.
Q: In practice, in large environments with multiple PetaBytes on
record in a single domain's catalog, using a mix of DLT 7000 and SDLT 220
tape drives and their media (two different animals - I know), to what
extent should a hard expiry date be de-emphasized?
Q: Is the default of 200 mounts a useful real-world end-of-life
measure for DLT or SDLT-I media?
Q: How many shops actually wait until individual tapes fail with
frequency and are FROZEN to retire tapes?
Here's why I ask. NetBackup v4.5 allows a tape to silently expire
within a library and become a ZOMBIE.
While informed administrative practice would defend against this
problem, (fill in my excuse...).
I recently had to address a problem in a large (500 tapes, 10 DLT
7000 drives) tape library with 48 ScratchPool tapes in it, yet none were
migrating into the other production Volume Pools when they ran low. Lots
of 96 errors.
After digging into the legacy scripts the Operations staff use
daily to identify how many ScratchPool tapes they need to inject into the
various libraries, vmquery was correctly reporting that this library did
have 48 ScratchPool taped loaded. One of these has the lowest value Media
ID in the entire Domain, so it was not much of a leap to the correct
conclusion. All were ejected, placed in a new EXPIRED Volume Pool, and
ScratchPool was properly loaded into the library.
Now I want to determine what to do with the expired volumes.
Q: Extended their use by tacking another 2 years on to their lives?
Q: If so, do I down-select their allowed mount counts?
Q: Or do I just wait until they fail?
I know I have suffered declining expectations with NetBackup
Enterprise v4.5 (scheduled to go v5.1 FP1 in November), especially in
recovery behaviour after FC-SCSI router "glitches" (Chicken Little would
use more enthusiastic word choice :-). While I'm reducing my ignorance on
the operational side of an NBU environment, I did not expect Volume Expiry
to be an unremarked event.
Q: On a day-to-day basis, how do large environments protect against
ZOMBIE behaviour of expired Volumes, especially when the clock strikes
midnight while the Volume is in a library?
--=_alternative 0070BA4986256F20_=
Content-Type: text/html; charset="us-ascii"
<br><font size=1 color=#800080 face="sans-serif">----- Forwarded by Michael F
Lavelle/LAKE/CHMS/ABBOTT on 10/01/2004 03:30 PM -----</font>
<br>
<table width=100%>
<tr valign=top>
<td>
<td><font size=1 face="sans-serif"><b>Michael F Lavelle</b></font>
<p><font size=1 face="sans-serif">10/01/2004 11:47 AM</font>
<br>
<td><font size=1 face="Arial"> </font>
<br><font size=1 face="sans-serif"> To:
veritas-bu-admin AT mailman.eng.auburn DOT edu</font>
<br><font size=1 face="sans-serif"> cc:
Michael F Lavelle/LAKE/CHMS/ABBOTT@ABBOTT</font>
<br><font size=1 face="sans-serif"> Subject:
Volume Expiration and ScratchPool</font></table>
<br>
<br><font size=2 face="Arial">Hi,</font>
<br><font size=2 face="Arial"> I don't want to start
a shoot-from-the-hip philosophical discussion, but my topic is Volume
Expiration.</font>
<br>
<br><font size=2 face="Arial"> I understand the
treatment of tape expiration varies widely.</font>
<br>
<br><font size=2 face="Arial">Q: In practice, in
large environments with multiple PetaBytes on record in a single domain's
catalog, using a mix of DLT 7000 and SDLT 220 tape drives and their media (two
different animals - I know), to what extent should a hard expiry date be
de-emphasized?</font>
<br><font size=2 face="Arial">Q: Is the default of
200 mounts a useful real-world end-of-life measure for DLT or SDLT-I
media?</font>
<br><font size=2 face="Arial">Q: How many shops
actually wait until individual tapes fail with frequency and are FROZEN to
retire tapes?</font>
<br>
<br>
<br><font size=2 face="Arial"> Here's why I ask.
NetBackup v4.5 allows a tape to silently expire within a library and
become a ZOMBIE.</font>
<br><font size=2 face="Arial"> While informed
administrative practice would defend against this problem, (fill in my
excuse...).</font>
<br>
<br><font size=2 face="Arial"> I recently had to
address a problem in a large (500 tapes, 10 DLT 7000 drives) tape library with
48 ScratchPool tapes in it, yet none were migrating into the other production
Volume Pools when they ran low. Lots of 96 errors.</font>
<br><font size=2 face="Arial"> After digging into
the legacy scripts the Operations staff use daily to identify how many
ScratchPool tapes they need to inject into the various libraries, vmquery was
correctly reporting that this library did have 48 ScratchPool taped loaded.
One of these has the lowest value Media ID in the entire Domain, so it
was not much of a leap to the correct conclusion. All were ejected,
placed in a new EXPIRED Volume Pool, and ScratchPool was properly loaded into
the library.</font>
<br>
<br><font size=2 face="Arial"> Now I want to
determine what to do with the expired volumes.</font>
<br>
<br><font size=2 face="Arial">Q: Extended their use
by tacking another 2 years on to their lives?</font>
<br><font size=2 face="Arial">Q: If so, do I
down-select their allowed mount counts?</font>
<br><font size=2 face="Arial">Q: Or do I just wait
until they fail?</font>
<br>
<br>
<br><font size=2 face="Arial"> I know I have
suffered declining expectations with NetBackup Enterprise v4.5 (scheduled to go
v5.1 FP1 in November), especially in recovery behaviour after FC-SCSI router
"glitches" (Chicken Little would use more enthusiastic word choice
:-). While I'm reducing my ignorance on the operational side of an NBU
environment, I did not expect Volume Expiry to be an unremarked event.</font>
<br>
<br>
<br><font size=2 face="Arial">Q: On a day-to-day
basis, how do large environments protect against ZOMBIE behaviour of expired
Volumes, especially when the clock strikes midnight while the Volume is in a
library?</font>
<br>
--=_alternative 0070BA4986256F20_=--
|
