Inventory robot on ACSLS library

[faridux]

Hello,

First I would like to thank you, any help will be appreciated
Here is a brieve description of the infra

I have two solaris Netbackup master:
- master1: V6.5.5, pool_id configured on the acsls:3000, the library is configured as ACS 2
- master 2: V6.0 MP7, pool_id configured on the acsls:2000, the library is configured as ACS 4

I have a lack of available tape on master2,
thus I have transferred on my acsls, 20 LTO4 tapes from pool 3000 to pool 2000 (with set scratch command)
Everything looks fine on the acsls as, with the volrpt command, these 20 tapes appeared to be now in the right pool (2000)
Nevertheless when I run an inventory robot on both master, it tells "Volume configuration is up-to-date with robot contents"

here is an extract of vm.conf on master 1:
master1:/usr/openv/volmgr# more /usr/openv/volmgr/vm.conf | grep INVENTORY
INVENTORY_FILTER = ACS 2 BY_ACS_POOL 3000

here is an extract of vm.conf on master 2 (this master is connected to 2 others acsls library 2 and 3 not involved):
toxsos002 # more /usr/openv/volmgr/vm.conf | grep INVENTORY
INVENTORY_FILTER = ACS 2 BY_ACS_POOL 2000
INVENTORY_FILTER = ACS 3 BY_ACS_POOL 2000
INVENTORY_FILTER = ACS 4 BY_ACS_POOL 2000

Is there anything I missed, or do I have to do any additionnal operation?
Thank you

[Marianne]

Confirm that NBU's query to ACS pool is well by using robtest:
Select robot number and type:
qpool 2000

and one of the new volumes:
qvol <vol>

Maybe also check volume ownership on ACSLS (if that was implemented).

[faridux]

I have just run a robtest on master 1, here is the output:

Robot Selection
---------------
  1)  TLD 0
  2)  TLD 1
  3)  ACS 2
  4)  none/quit
Enter choice: 3

Robot selected: ACS(2)   ACSLS host = acsls5, SSI socket = 13741

Invoking robotic test utility:
/usr/openv/volmgr/bin/acstest -rn 2 -r acsls5 -s 13741

Server 0 with 1728 free cells is in state "STATE_RUN"

QUERY SERVER complete

Enter acs commands (? returns help information)
qpool 3000
Pool ID 3000 has 139 volumes
QUERY POOL complete

q vol B02270
(When I type [Enter] it has exited to the main menu hereunder, I guess it is normal, as it is not  supposed to access to it anymore)

Robot Selection
---------------
  1)  TLD 0
  2)  TLD 1
  3)  ACS 2
  4)  none/quit
Enter choice: nbumtr1:/usr/openv/volmgr# robtest
Configured robots with local control supporting test utilities:
  TLD(0)     robotic path = /dev/sg/c0tw518a90543f9d8010l0
  TLD(1)     robotic path = /dev/sg/c0tw518a9056f0438010l0
  ACS(2)     ACSLS host = acsls5

Here is below the output on master2:
toxsos002 # robtest
Configured robots with local control supporting test utilities:
  ACS(3)     ACSLS host = acsls5
  ACS(4)     ACSLS host = acsls5
 

Robot Selection
---------------
  1)  ACS 3
  2)  ACS 4
  3)  none/quit
Enter choice: 2

Robot selected: ACS(4)   ACSLS host = acsls5, SSI socket = 13741

Invoking robotic test utility:
/usr/openv/volmgr/bin/acstest -rn 4 -r acsls5 -s 13741

Server 0 with 1728 free cells is in state "STATE_RUN"

QUERY SERVER complete

Enter acs commands (? returns help information)
qpool 2000
Pool ID 2000 has 6 volumes
QUERY POOL complete

qvol B02270
B02270   LTO_800G   home       2, 1, 14, 5, 1         
QUERY VOLUME complete

Then I exit from robtest,
I guess this means the qvol works fine, as acsls replies to master2 with volume location, and deny the request of master1
Moreover, no event.log was created in /usr/openv/volmgr/debug/acsssi/

Thank you for your assistance


 

[Kevin_Good]

...
q vol B02270

(When I type [Enter] it has exited to the main menu hereunder, I guess it is normal, as it is not  supposed to access to it anymore)

...

robtest picked up 'q' for quit
you meant qvol for query volume

[Marianne]

Kevin is correct - there should be no space in 'qvol'. I was hoping to see more info with qvol, i.e. something like media ownership (haven't worked with these commands in a while...).

You say that you have transferred 20 tapes to pool 2000, but qpool says that there are only 6 tapes in this pool:
qpool 2000
Pool ID 2000 has 6 volumes

If you query pool 2000 from ACSSA>  it will probably display the same output - 6 tapes?
Something probably wrong with the way that you 'transferred' from one pool to another?

If you share the exact commands that you've used in ACSLS, I can ask my colleagues doing ACSLS support to double-check them for you.

[faridux]

Thank you Kevin and Marianne
Yes you are right, I did a mistake, sorry hereunder is the output, the acsls replies also to master1:

Robot Selection
---------------
  1)  TLD 0
  2)  TLD 1
  3)  ACS 2
  4)  none/quit
Enter choice: 3

Robot selected: ACS(2)   ACSLS host = acsls5, SSI socket = 13741

Invoking robotic test utility:
/usr/openv/volmgr/bin/acstest -rn 2 -r acsls5 -s 13741

Server 0 with 1728 free cells is in state "STATE_RUN"

QUERY SERVER complete

Enter acs commands (? returns help information)
qvol B02270
B02270   LTO-800G   home       2, 1, 14, 5, 1         
QUERY VOLUME complete

q

Robot Selection
---------------
  1)  TLD 0
  2)  TLD 1
  3)  ACS 2
  4)  none/quit
Enter choice: nbumtr1:/usr/openv# robtest
Configured robots with local control supporting test utilities:
  TLD(0)     robotic path = /dev/sg/c0tw518a90543f9d8010l0
  TLD(1)     robotic path = /dev/sg/c0tw518a9056f0438010l0
  ACS(2)     ACSLS host = acsls5

About acsls qpool command,you are right Marianne there is something wrong,
On Netbackup side on master2
- the other tapes already known, they are mounted and used correctly I guess EMM is ok,
- hereunder the number of know LTO4 on master2:
toxsos002 # /usr/openv/netbackup/bin/goodies/available_media | grep "^B" | wc -l
     241
- On the acsls software I can see there are 261 (241 + 20 I have transferred) LTO4 in the pool 2000 of master2:
bash-3.00$ volrpt -a 2 -f cartouches.txt | grep 2000 | head -1
B00900  2000   D    
bash-3.00$ volrpt -a 2 -f cartouches.txt | awk '{print $2}' | grep 2000 | wc -l
     261
I have only one technote about acsls http://seer.entsupport.symantec.com/docs/267306.htm
I keep searching too about the uncorrect output of the qpool command,
Thank you for your help

[Marianne]

Please run 'q pool 2000' from ACSSA> prompt on acsls server (instead of volrpt)

[faridux]

Sorry, here is the output of the command
ACSSA> q pool 2000
2010-06-30 11:07:10               Pool Status
 Identifier  Volume Count  Low Water Mark   High Water Mark  Attributes
 2000        7             0                2147483647    

I think this highlights something wrong in the acsls configuration, isn't it?

[Marianne · Accepted solution]

My thought exactly. All troubleshooting must now be done on ACSLS.

Extracts from ACSLS manual:


To confirm specific media-id's belonging to each pool, use q scratch <pool_id>
 to list media-id's.

Note: Scratch volumes with a volume status of absent, ejected, or missing are not included in the Query Scratch output.


Rebalancing Scratch Pools
Use this procedure to rebalance scratch pools by moving scratch volumes from one pool to another.
To rebalance scratch pools, do the following:
1. To display the attributes of all scratch pools, enter:
query pool all
For more information see “query pool” on page 378
2. Use the query scratch command to display the IDs of scratch volumes in pools you want to rebalance.
For more information, see “query scratch” on page 382
3. Use the set scratch command to move scratch volumes from one pool to another.
For example, to move volumes YUMA20 through YUMA80 (which currently reside in pool 5) to pool 10, enter:
set scratch 10 YUMA20-YUMA80
For more information, see “set scratch” on page 395


(Information extracted from ACSLS 7.1 Installation, Configuration, and Administration Guide)

[faridux]

Your comment is very helpfull, as this has allowed me to see I have an inconsistency between the information I get with volrpt command and the q scratch in the cmd_proc
I will open a call to Sun 
Nevertheless for my initial problem maybe I found something,
I have transferred the 20 tapes from pool 3000 to pool 0 with the command: set scratch off ,then to pool 2000

because of the off I used in my command the VOLUME_TYPE was still D (for data) instead of S for scratch

here is what I have just done, maybe this can help someone else:

I have deleted them from master1 (with the vmdelete command)
I put them back to pool 0 with set scratch command (without the off!)
The flag attribute VOLUME_TYPE has changed from D to S
then I moved the tapes from pool 0 to pool 2000
Finally the master2 has detected the tapes during the inventory

Thank you again for your help Marianne, I mark your post as solution as the problem is solved,
Nevertheless I will follow the q scratch problem with my Sun support
Regards