cancel
Showing results for 
Search instead for 
Did you mean: 

Multiple libraries controlled by single acsls cluster

Dr_Strangelove
Level 3
Hi,

My environment is running nbu 6.5.3 and a second SL8500 has recently been added. This new library is controlled by the same ACSLS cluster controlling an existing SL8500, so this new library has become ACS1 on the ACSLS cluster.

What I am wondering is what must I do within NBU to configure this unit. A robot is already created and pointing at the ACSLS cluster for the existing SL8500 so presumably I do not need to create another robot for the new library as ACSLS will know by the drive addresses what is destined for ACS 0 and what is destined for ACS 1.

No additional scratch pools have been created so at the moment the only scratch pool is the default 0. When an inventory is run from NBU on the ACS robot it only picks up the existing library and not the new library. Is it the case that I need to create a new scratch pool for the new SL8500 and put the relevant ACS inventory filter in the vm.conf or do I need to do anything with the robotic configuration to enable my NBU servers to run an inventory on the new library ?

Thanks
7 REPLIES 7

Nicolai
Moderator
Moderator
Partner    VIP   
Are there pass thru ports (PTP's) installed in order for the two SL8500 to exchange tapes ?

Dr_Strangelove
Level 3
There are no passthru ports installed, the two libraries are standalone units.

Deepak_W
Level 6
Partner Accredited
If these are two different TL units having two different robots and different drives, then there must be two different robots listed in your NBU master.

You get the same robot for the TL when those TLs share same robot and use pass through port. As mentioned by Nicolai Pass through port helps you exchange the tapes.

So if you have two TLs then make sure you get two seperate robots and also you'll need to configure two seperate Storage Units for two libraries.

Check how many robots are visible to NBU using scan command. You can run this command from volmgr directory.

scan command will display all visible robots and drives.


So thinking from your side, you will need to create new storage unit for new TL and will need to assign the same to the policies which you want to use this new TL.


Nicolai
Moderator
Moderator
Partner    VIP   

NO NO - You can't use the scan command for ACSLS robots - They are API controlled. Drive will however be visible.  Treat the two SL8500  as separate robot's   - and don't get confused by ACSLS control them both.  

Are there a reason for using ACSLS scratch pools and  inventorying filters ?




Omar_Villa
Level 6
Employee

Any new ACS it must be configured has a totaly different and new silo, doesnt matter who will control that, the ACSLS box is only the library controler.

Dr_Strangelove
Level 3
Thanks for the help so far..

How does NetBackup differentiate between the two silos however as when the ACS robot is created within NBU you only specify the robot control host which is the ACSLS cluster virtual hostname.  There is nothing else I am aware of which can tell NBU that you want to ACS 1 rather than ACS 0.

As I have not configured such a setup before I was under the assumption that ACSLS would treat the 2 libraries as one logical unit that was why I was thinking maybe scratch pools and inventory filters were required to enable the volumes in the default pools on both libraries to be visible to NBU. From cmd_proc and acstest etc the 2 ACS units appear as one logical entity, i.e. volume queries and drive queries return the volumes and drives from both libraries by default.



Nicolai
Moderator
Moderator
Partner    VIP   
You have a lot of questions, I have tried to answer them all - Please raise a hand if you are still left with a question.

ACSLS is able to control a large amount of STK silos but luckily Netbackup has a great integration with ACSLS .  When performing a Netbackup robot  inventory, ACSLS return the location (ACS,LSM,panel,row,column) of the  tapes along with it type (HCART,9840,LTO etc etc). This way Netbackup (and ACSLS) ensures that tapes are not mounted across non' PTP'ed robots and a 9840 tapes isn't inserted into a LTO tape drive.

Only robot that are PTP'ed will appear as one logical unit (the ACS will be the same but you will have multiple LSM's just like today).  cmd_proc is a ACSLS tool and can always see all robots,tapes and drives, but that does not mean the mapping in Netbackup is the same. Query from Netbackup also return info  from both robot because ACSSL control them both.

You don't need inventorying filter if it's the same backup system that access both SL8500. If it's different system,  use ACSLS volume access control.