01-12-2011 09:04 AM
Hey
I've got a challenge with one of my backup environments that someone may have a quick answer to!
master: Solaris, 1 x TS3310 Robot and 3 FC LTO4 drives
media : AIX 6.0, 2 x FC LTO 4 Drives in above robot
When a backup is initiated using the media server, a tape request is called an passed to the robot. The robot then mounts the tape in the drive (confirmed by Robtest and via the web interface on the 3310).
However, confirmation of this mount isn't passed back to the running job which eventually fails and downs the drive with an extended media error.
The drives are visibile to the o/s via "lsdev -Cc tape" and have been reconfigured along with recycling netbackup.
scan/tpautoconf -t shows the drives aswell.
Help!!
Thanks
01-12-2011 10:02 AM
First check we got the right drives on the os, to the correct drives in Netbackup.
Go into the robot gui and look at the serial numbers for the drives.
Start with Module 0 drive 0,1 (top drive in the module)
Physical SN: | 1310006660 |
Now you need to make sure that they are matching correctly in netbackup
( I had to figure this out when I first got the TS3310)
The drive numbers for the robot for the TS3310 work like this
Module 0 -
drive 0,1 - robot drive 5
drive 0,2 - robot drive 6
Module -1
drive -1,1 - robot drive 1
drive -1,2 - robot drive 2
drive -1,3 - robot drive 3
drive -1,4 - robot drive4
-----
the serial number of the drive you see in the robot gui, should be matching the same serial number to the robot drive number you have configured in the nb gui. You can see this in devices/drives with the column Serial Number.
if they do not match this is where your issue is. The wizard should be able to configure this correctly (I have aways done mine by hand to make sure they are)
-----
What I think is happening, is the nb is putting the tape in robot drive 1 (which say is rmt6) and aix is looking for the tape in rmt1 which is a different robot drive number.
I don't know how many modules or drives you have so I could only give an example.
01-12-2011 12:08 PM
thanks for the reply.
The device serials match in the robot and Netbackup. This setup has been working successfully for months and failed at some point over the weekend.
BPTM wasn't logging so we switched that on today and found this when a tape mount is attempted:
reserve unit scsi command failed, status = 0x2, key = 0x2, asc = 0x3e, ascq = 0x0
I'm not sure what this means, but I hope its a clue as to whats wrong.
01-12-2011 12:25 PM
Do you have more then on media server and are you using SSO.
It sound like the drive is locked on another reservation
01-12-2011 12:59 PM
No SSO, all drives are direct attached.
1 Master with the robot control and 3 drives, and the the media with 2 drives.
01-12-2011 01:10 PM
The last thing I can think of is to verify on AIX that the tape drive is still avail and not just defined.
01-12-2011 01:19 PM
They are indeed - lsdev -Cc tape shows them as available :(
Its really frustrating!!!
01-12-2011 02:33 PM
Could you post the lsdev -Cc tape output?
Thanks.