Thanks Michael
I have learned something new today…
This is what I’ve done:
I
went through the cleanup process (done many times before):
Stop all services on master & media servers.
Delete globDB ltidevs & robotic_def on master & media
Start vmd on all
run ' vmglob -set_gdbhost <master>' on all
Run device wizard.
The
wizard picked up and added the 13 drives seen by Master and Media, all fine,
BUT vmglob -listall still shows 6 orphaned entries...
I've
been racking my brain to try and figure out where these orphaned entries are
coming from - Seems these drives are still physically in the robot - a total of
19 drives. The 6 drives that show up in vmglob output with only a serial
# can be seen by the 'scan' command when the robot reports on drives
physically in the robot.
The
highlighted drives are not connected/zoned to any server. Drive bay 5 is empty.
Drive 1 Serial
Number : "PXB15P1398"
Drive 2 Serial
Number : "PXB15P2419"
Drive 3 Serial
Number : "HM005KV"
Drive 4 Serial
Number : "HM003A6"
Drive 5 Serial Number
: ""
Drive 6 Serial
Number : "HM005K1"
Drive 7 Serial
Number : "HS00VGZ"
Drive 8 Serial
Number : "HM0057K"
Drive 9 Serial
Number : "HM008L3"
Drive 10 Serial
Number : "HS00P8K"
Drive 11 Serial
Number : "HU10527857"
Drive 12 Serial
Number : "HU10609UL2"
Drive 13 Serial
Number : "HU105279E4"
Drive 14 Serial
Number : "HU10616BRY"
Drive 15 Serial
Number : "HU1052784V"
Drive 16 Serial
Number : "HU1052784K"
Drive 17 Serial
Number : "HU1052798A"
Drive 18 Serial
Number : "HU1052798M"
Drive 19 Serial
Number : "HU1052796W"
Drive 20 Serial
Number : "HU1052798E"
These
drives were deleted from NBU a long time ago. Seems the device wizard still
adds these tape drives (reported by the robot) to the globDB, although none of
the media servers can see it at O/S level.
From: Michael Graff Andersen
[mailto:mian71 AT gmail DOT com]
Sent: 06 August 2010 12:21 PM
To: Marianne Van Den Berg
Cc: veritas-bu AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu] Orphaned globDB entries
Think the tpautoconf command, may able to help you
2010/8/6 Marianne Van Den Berg <mvdberg AT stortech.co DOT za>
Hi
all
Hope
there’s someone out there with a quick answer…
We
are about to upgrade a 5.1 Solaris master to 6.5.
nbcc
was submitted to Symantec, all inconsistencies cleaned up.
I
ran vmglob –listall to create output of device config and found
some old, orphaned tape drive entries in globDB with only serial number (no
device or host name).
We
tried to delete the entry with ‘vmglob –delete –serial
HU1052784L ‘. The command completed
successfully, but the entries are still there:
#vmglob -listall -b
device
type device
name serial
number
host name
------------------------------------------------------------------------
drive HPUltrium3-SCSI3
HU10616BRY
<master>
drive
HPUltrium3-SCSI
HU1052784K
<media>
drive
-
HU1052784L
-
drive
-
HM008L3
-
drive -
HM0057K
-
drive
-
HM005K1
-
drive
-
HM0077K
-
drive
-
HM003A6
-
drive
-
PXB15P2419
-
drive
-
PXB15P1398
-
Does
anyone have any idea how to clean this up (preferably without having to delete
and re-create all device entries)!
Our
case engineer at Symantec seems to be ‘missing in action’.
Regards
M.
_______________________________________________
Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu