NetBackup Finds Library, But Not Drives Attached using NDMP

[Porkins]

Running NetBackup 7.1 on Windows Server 2008 R2. I have an HP StorageWorks MSL4048 tape library with 2 drives installed. These are attached with fibre to 2 fibre channel ports on the back of my NetApp FAS2240. When I run sysconfig -t and sysconfig -m on the NetApp, it shows the 2 tape drives, and the library. When I try to discover these devices in NetBackup using the storage wizard, it only finds the robotic library, and lists the 2 drives as "unused element address". I have seen this happen once before, but that was a probelm with an old NetApp that did not have tape config files for the newer LTO5 drives. This is a new NetApp that already has the HP and IBM LTO5 drive tape config files. And the NetApp detects them successfully anyway. Any ideas why NetBackup does not? We verified that ndmp is ON, and the account name and password should be good, otherwise it would detect nothing at all. I can even unplug the fibre, and run the wizard again, and it detects nothing, so I know that is all working correctly.
 

[Yasuhisa_Ishika]

"unused element address" means that tape library reported drives but NetBackup could not detect drives. Can you see any drives in "toautoconf -probe <your_filer>"?

[SymTerry]

A couple of things can cause an unused element address error. Usually robot issues and tape drivers. Have you tried to uninstall and reinstall the tape drivers in windows?

[Abesama]

Hi,

With fibre (same for plain SCSI) tape library/drive detection issues, this sometimes helps.

1. Power off all devices
2. Ensure cable is all connected
3. Ensure tape drives do not have any tape loaded, media access port and transport arm are all empty
4. Turn on the tape library, give it few minutes so that it will be ready to respond to server
5. Turn on the NetApp (if it was not an NDMP scenario then media server) so that it can freshly detect tape devices and create the new nrst paths
6. Scan again from the NetBackup GUI wizard or run tpautoconf command syntax

That being said, if it were not HP MSL, some LUN masking work might be required from the tape library's admin console GUI but in your case I don't think that is necessary.

Hope this helps, good luck.

A

[Marianne]

Unused element is normally the robot itself thinking there are no drives installed.
This has to do with control path installation - a connection between the robot and the tape drives - not the fibre connection that goes to the switch/servers.

Is the robot connected to the Windows server as well? If so, run 'scan -changer ' to show us what the robot sees internally.

Please also double-check that up to date NBU device mappings are installed on the Windows server. I will post the URL a bit later.

*** Update ***

Device mappings URL: http://www.symantec.com/docs/TECH227191

[sdo]

Are the two drives connected as one tape drive per FAS controller head?

In such a situation, one of the NetApp controller heads will probably see the robot control LUN, and also the LUN of the first tape drive - and the second NetApp controller head will see one tape drive LUN, but no robot control LUN.

You may have to partition your tape library, so that each FAS controller head can each see one robot control LUN, and one tape drive LUN.

[Porkins]

Running "tpautoconf -probe sam" returns:

Host "sam" SCSI device model "HP MSL G3 Series"

It gives me the WWN, Serial, Electrical Name, and Alias 0=mc0.

It does not return any drives.

If I connect the tape library to the server directly, Windows discovers and installs drivers for the Medium Changer, and for the 2 tape drives. If I run "scan -changer", it returns the correct information, showing "Number of Drives:2", and the serial numbers for both drives.

[sdo]

Is both tape library firmware and tape drive firmware up to date? Is the FAS2240 a single unit - or two heads? Are both NetApp fibre ports in initiator mode? Have you confirmed HCL each way - I.e check HP MSL4048 to see if they support FAS2240, and check FAS2240 to see if they support HP MSL4048?

[Yasuhisa_Ishika]

From NetBackup perspective, NetApp dies not show any tape drives through NDMP.

2 fibre channel ports - does it mean each drives are connected to 2 ports of single controller(multipath)? What version and mode of Data ONTAP are you using? Can you podt actual output of "sysconfig -a", "sysconfig -t", "sysconfig -m", "storage alias", "storage show mc", and "storage show tape"?

[Porkins]

It looks like the NetApp is not able to correctly identify the tape drive. Here is what I see:

sysconfig-a shows me the medium changer and the 2 tape drives connected to the fibre ports 1a and 1b. It shows Tape: HP Ultrium 5-SCSI Y5BW for both.

sysconfig -t shows maximum number of Dynamic Tape Qualification configuration files (64) has been exceeded. Tape Drive (1b.125) HP Ultrium 5-SCSI (Non-qualified tape drive) and Tape Drive (1a.125) HP Ultrium 5-SCSI (Non-qualified tape drive). I tried to delete some of the TCF files for drives that I know will NEVER be used, but I still see this error.

storage show tape gives me error: unable to obtain info for device <unknown>  Device State: available

When I unplug the fibre, wait 1 minute, then reconnect the fibre while I have an SSH window opened up and connected, I see this error message: Data ONTAP does not recognize this tape drive: 1b.125: HP Ultirum 5-SCSI Y5BW.

I double-checked the etc/tape_config folder, and there are a bunch of tcf files in there, including the HP_LTO5_ULTIRUM.TCF file, which I assume is the one it wants to use. Is a NetApp reboot in order? Is there something I can do to eliminate the number of Dynamic Tape Qualification configuration files error?

[Yasuhisa_Ishika · Accepted solution]

Have a look at Burt#114013. Too many tape device definition is loaded. Remove all the files except you really needed, and reboot the controller. For futher assistance, contact NetApp support.

http://mysupport.netapp.com/NOW/cgi-bin/bol?Type=Detail&Display=114013

[Porkins]

I removed all excess TCF files and rebooted. Everything works well now! Thank you all!