Unsolved
This post is more than 5 years old
2 Posts
0
2211
Inquire on Solaris 10 showing duplicate entries for same tape device
Well I thought I would start here. First off, don't know Networker what's so ever. I work for HP and support the virtual Tape library being presented to the the Networker SAN servers.
In some cases, the inquire command will return duplicate device for the same tape device, or 2 entries for the same device, Same LUN and /dev/rmt file, just a different target ID.
The Fiber HBA's are Emulex 8 GB HBA's and they are using the native Solaris fiber driver, (Leadville). The problem can occur on any one of the Networker SAN backup servers.
Below is an example:
WWNN=62AF9B49DB680060200000000000C0D5
scsidev@4.8.0:HP Ultrium 3-SCSI R210|Tape, /dev/rmt/408cbn
S/N: d9CW.ORs0v
ATNN=HP Ultrium 3-SCSI d9CW.ORs0v
WWNN=62AF9B49DB6800601F0000000000C0D5
scsidev@4.10.0:HP Ultrium 3-SCSI R210|Tape, /dev/rmt/410cbn
S/N: d9CW.ORs8v
ATNN=HP Ultrium 3-SCSI d9CW.ORs8v
WWNN=62AF9B49DB6800601F0000000002C0D5
scsidev@4.12.0:HP Ultrium 3-SCSI R210|Tape, /dev/rmt/410cbn
S/N: d9CW.ORs8v
ATNN=HP Ultrium 3-SCSI d9CW.ORs8v
WWNN=62AF9B49DB6800601F0000000002C0D5
Notice that the last 2 devices are really the same device/
cfgadm -alv -o show_FCP_dev show correct number of devices.
Any ideas?
-Julian
tlemons1
87 Posts
0
September 9th, 2011 13:00
Hi Julian
Please post the driver version and the FW version in the Emulex card.
Thanks
tl
nearlinegeek
2 Posts
0
September 9th, 2011 14:00
HBA Port WWN: 10000000c99774a5
OS Device Name: /dev/cfg/c3
Manufacturer: Emulex
Model: LPe11002-S
Firmware Version: 2.80x7 (Z3F2.80X7)
FCode/BIOS Version: Boot:5.02a1 Fcode:1.50a9
Serial Number: 0999BT0-1001000A4M
Driver Name: emlxs
Driver Version: 2.40s (2009.07.17.10.15)
Type: N-port
State: online
Supported Speeds: 1Gb 2Gb 4Gb
Current Speed: 4Gb
Node WWN: 20000000c99774a5
ble1
14.3K Posts
0
September 16th, 2011 13:00
Are you using some sort of multipathing? To check if those are two entries (really) for the same drive, test them with mt. I haven't work with Solaris much lately, but I remember that in old days numbering used to star from 0. In your case is 400+. Unless you use custom names (like /dev/rmt/{bus}{target}cbn then I wouldn't be surprised if device database might have gone banana. Without any further test to test few potential scenarios, it would be hard to say what it is. I would start by testing if I get the response on both device names if case I load the tape. If yes, it means both device names are valid and this might be due to multipathing.
--
Blog: https://community.emc.com/people/ble/blog
Thierry101
326 Posts
0
November 28th, 2011 18:00
hi nearlinegeek
have you checked the zoning? also, do you do port or node zoning?