Unable to discover drives after relocating master server to new site

[morbid-BMX-Co]

Hello Netbackup gurus/wonderkids,

I recently started a discussion in relation to fully decommissioning a site in Amsterdam and relocating the master server to a new site in the UK but only to be used for historic restores. That part of the process went well, thanks to help from some of you. The issue I face now is the master server has been setup in the new data centre but I am having trouble seeing the library or tape drives.

I have confirmed with my UNIX resource that the drives are visible at O/S level and she provided me with the path which is /dev/rmt/0. I have read quite a number of tech notes in relation to discovering tape drives but I have tried a few of these ideas but had no success so I though I could borrow the knowledge of you trusted folks.

I ran a couple of commands to see what NBU can see and got the following:

amsnbuma02:bin root # ./tpext -get_dev_mappings_ver
device mappings version in the EMM database is 1.78
device mappings version from the local file is 1.78
Local device mappings file is up-to-date

amsnbuma02:bin root # ./tpautoconf -t
amsnbuma02:bin root #

amsnbuma02:bin root # ./tpautoconf -report_disc
amsnbuma02:bin root #

I cant help thinking that I have missed a fundemental step before trying these commands but Im not sure what else to do so any help you can offer is greatly appreciated. This is the first time have attempted anthing like this and only had a years experience in NBU so still what you would call a 'noob'.

Also worth noting that we are running NBU 6.5.5 (I know its well out of date and support but this is only to be used for restores and Sky are reluctant to spend money on kit that will be redundant in the next couple of years and O/S of the master server is SunOS 5.1

[morbid-BMX-Co]

Attempted robtest too and got the following - No robots are configured

[Pritesh_Pisal]

How the library connected to the server physically or it’s an SAN? I would suggest kindly check the library connectivity part.if robtest is not able to view robot, then there is a problem with either connectivity or HW.

[revarooo]

Can you run:

/usr/openv/volmgr/bin/scan

What is the output of that? This will tell us what the OS can actually see.

[Marianne]

I have confirmed with my UNIX resource that the drives are visible at O/S level and she provided me with the path which is /dev/rmt/0.

Only one tape drive?  What about a tape library?

NBU uses the 'no-rewind' device name. So, it should be /dev/rmt/0cbn.

To ensure that the OS is seeing device correctly, show us output of

mt -f /dev/rmt/0cbn status

and

/usr/openv/volmgr/bin/sgscan

If OS sees device and sgscan not, you will need to rebuild sg driver. 
This is how I do it: recreate /dev/sg/* on solaris 10 without reboot 

If all is fine and sgscan reports tape drive(s) and robot, you can start the Device Config wizard.
 

[morbid-BMX-Co]

its all scsi connections Pritesh DCMT have confirmed all is connected but I know that doesnt mean everything.

revaroo - $ ./scan
************************************************************
*********************** SDT_TAPE    ************************
*********************** SDT_CHANGER ************************
*********************** SDT_OPTICAL ************************
************************************************************
 

[morbid-BMX-Co]

Marianne the command outputs are as follows:

$ mt -f /dev/rmt/0cbn status

/dev/rmt/0cbn: no tape loaded or drive offline

sgscan returns nothing

I did say to my UNIX resource there should be a library present too but as you described we can only see a tape drive.

Also worth noting Marianne that I marked one of your comments as a solution to the decom forum I started so thanks again.

[Marianne]

mt -f shows expected output.

Please follow the steps to rebuild the sg driver (see the link in my previous post) and run sgscan again.

[morbid-BMX-Co]

Thanks for that Marianne although the link below from your post didn't work to determine the number of targets and luns unfortunately.

*******See http://www.symantec.com/docs/TECH10725 for examples how to determine number of targets (to be used with -mt) and number of luns (-ml).*******

I have googled similar queries but not finding much. They all say to run the scan -changer and scan -tape commands which I cant run.

Any suggestions?

[morbid-BMX-Co]

# /usr/openv/volmgr/bin/driver/sg.install

Copied files to /kernel/drv to /kernel/drv/sparcv9.

Doing add_drv of the sg driver
Creating /dev/sg.
Added entry in /etc/devlink.tab file.
Made links in /dev/sg

/usr/openv/volmgr/bin/sgscan
/dev/sg/c0tw21000014c3cab01fl0: Disk (/dev/rdsk/c1t0d0): "SEAGATE ST314670FSUN146G"
/dev/sg/c0tw21000014c3cac37bl0: Disk (/dev/rdsk/c1t1d0): "SEAGATE ST314670FSUN146G"
/dev/sg/c0tw50017380ea7b0140l0: Array-controller: "IBM 2810XIV-LUN-0"
/dev/sg/c0tw50017380ea7b0140l1: Disk (/dev/rdsk/c2t50017380EA7B0140d1): "IBM 2810XIV"
/dev/sg/c0tw50017380ea7b0150l0: Array-controller: "IBM 2810XIV-LUN-0"
/dev/sg/c0tw50017380ea7b0150l1: Disk (/dev/rdsk/c2t50017380EA7B0150d1): "IBM 2810XIV"
/dev/sg/c0tw50017380ea7b0172l0: Array-controller: "IBM 2810XIV-LUN-0"
/dev/sg/c0tw50017380ea7b0172l1: Disk (/dev/rdsk/c3t50017380EA7B0172d1): "IBM 2810XIV"
/dev/sg/c0tw50017380ea7b0182l0: Array-controller: "IBM 2810XIV-LUN-0"
/dev/sg/c0tw50017380ea7b0182l1: Disk (/dev/rdsk/c3t50017380EA7B0182d1): "IBM 2810XIV"
/dev/sg/c0tw500507630f79700al0: Tape (/dev/rmt/0): "IBM ULT3580-TD4"
/dev/sg/c0tw500507630f79700al1: Changer: "IBM 03584L32"

So the sgscan shows an output now as you can see from above but when I run the scan command I still only get the below:

************************************************************
*********************** SDT_TAPE    ************************
*********************** SDT_CHANGER ************************
*********************** SDT_OPTICAL ************************
************************************************************
 

[Marianne · Accepted solution]

Seems Symantec has removed that TN in the meantime... No idea why...

You need to find out more about the actual tape library - is there only one tape drive installed?

If so, the OS single device name is fine and you have only one target for the tape drives.
Confirm by doing the following:

# cd /dev/rmt

# ls -l *cbn

If the above shows 0cbn as only output, use -mt 1 -ml 1.

This will create 2 targets for the sg driver (0 and 1) and the same for amount of luns per target.

Another option is to use sg.build just wil 'all' option as per Martin's reply to my post in the link above:

# /usr/openv/volmgr/bin/sg.build all

So, do the following:

# cd /usr/openv/volmgr/bin/driver
# /usr/openv/volmgr/bin/sg.build all 

Install the new sg driver configuration:

# /usr/bin/rm -f /kernel/drv/sg.conf
# /usr/openv/volmgr/bin/driver/sg.install

Check/verify config:

# /usr/openv/volmgr/bin/sgscan

Show us output of sgscan.

[morbid-BMX-Co]

right I did the wrong parameters originally so I have removed all files and ran through the process again so sgscan now shows:

amsnbuma02:driver root # /usr/openv/volmgr/bin/sgscan
/dev/sg/c0tw21000014c3cab01fl0: Disk (/dev/rdsk/c1t0d0): "SEAGATE ST314670FSUN146G"
/dev/sg/c0tw21000014c3cac37bl0: Disk (/dev/rdsk/c1t1d0): "SEAGATE ST314670FSUN146G"
/dev/sg/c0tw50017380ea7b0140l0: Array-controller: "IBM 2810XIV-LUN-0"
/dev/sg/c0tw50017380ea7b0140l1: Disk (/dev/rdsk/c2t50017380EA7B0140d1): "IBM 2810XIV"
/dev/sg/c0tw50017380ea7b0150l0: Array-controller: "IBM 2810XIV-LUN-0"
/dev/sg/c0tw50017380ea7b0150l1: Disk (/dev/rdsk/c2t50017380EA7B0150d1): "IBM 2810XIV"
/dev/sg/c0tw50017380ea7b0172l0: Array-controller: "IBM 2810XIV-LUN-0"
/dev/sg/c0tw50017380ea7b0172l1: Disk (/dev/rdsk/c3t50017380EA7B0172d1): "IBM 2810XIV"
/dev/sg/c0tw50017380ea7b0182l0: Array-controller: "IBM 2810XIV-LUN-0"
/dev/sg/c0tw50017380ea7b0182l1: Disk (/dev/rdsk/c3t50017380EA7B0182d1): "IBM 2810XIV"
/dev/sg/c0tw500507630f79700al0: Tape (/dev/rmt/0): "IBM ULT3580-TD4"
/dev/sg/c0tw500507630f79700al1: Changer: "IBM 03584L32"

And to answer an earlier question from the output above Im sure we only have one drive as this environment is ONLY to be used for restores so Im pretty sure its just going to sit there until everything has expired.

[morbid-BMX-Co]

Just want to say a massive thanks to you all for your valued input especially Marianne. Its comforting to know that if you put the question out there the response you get is usual quick and accurate! I can now see the library and drive and got everything back working as it should.

Cheers again folks

[Marianne]

Great stuff!

You have no idea how stumped I was the first time I experienced a situation where the OS could see devices on Solaris but NBU could not. And a customer staring at me because NBU is not working!
I had no access to Internet and only the text version of Device Config Guide in /usr/openv/volmgr (wonder if it is still there).
So, this is where I stumbled upon sg.install and sg.build. It took me a while to realize that the manual had these 2 commands the wrong way round and that sg.build should be done before sg.install....

[morbid-BMX-Co]

The situation you described sounds very uncomfortable indeed Marianne! But great effort in finding the solution in the text version. I did notice these commands whilst reading some tech notes but being fairly new to NBU I was unsure the way to go about it. Sometimes its just nice to get confirmation from someone who has performed the actions previously.

We have been kind of left to our own devices to sort things and when you delve into the world of the unknown reassurance is required so thanks for your patience and advice.

I will no doubt raise another issue if and when I stumble accross it.