Forum Discussion

Iwan_Tamimi's avatar
Iwan_Tamimi
Level 6
11 years ago

Problem setting up Netbackup SSO (Shared Storage Options)

Hi All,

 Our Netbackup system is version 7.5.0.4. We have 2 tape library (HP ESL322e) connect using SAN. We already had up running drives running as SSO (setup by our System Integrator). Two of  drives on the TL being shared between 2 server , one running Solaris 10 the other one AIX 6.1. in site A. There is another pair with the similar configuration on site B.

One day our HP did firmware upgrade to TL on site A. The upgrade was not smooth, I had to reconfigure again the connection on Solaris 10 (but the AIX was fine).

The problem if I setup individual tape drive to the Solaris  or AIX, the backup could run fine. Then I did the SSO with configuration like 1 only for solaris, another 1 only for AIX, but 2 other dirves I config as SSO. Then I test the backup on Solaris , it supposed can see 2 drives (1 single, 1SSO). I fired first backup it run on the Solaris, but when I run the second drive, it won' go that SSO drive, it keep on waiting for resources.

 

10/23/2013 16:17:16 - awaiting resource ESL025_EBS6_MPX. Waiting for resources.
          Reason: Drives are in use, Media server: ebs6-bck,
          Robot Type(Number): TLD(1), Media ID: N/A, Drive Name: N/A,
          Volume Pool: test1, Storage Unit: ESL025_EBS6_MPX, Drive Scan Host: N/A,
          Disk Pool: N/A, Disk Volume: N/A

notes: ebs6 is the Solaris media server.

 

What i did wrong? Why it cannot use that SSO drive?

 

Thank you and regards,

 

Iwan

7.5
Backing Up
Backup and Recovery
NetBackup
Restore
  • Marianne's avatar
    Marianne
    11 years ago

    Thanks for the output.

    Firstly, we see that stu ESL025_EBS6_MPX  is configured for 4 drives:

    Label:                ESL025_EBS6_MPX
Storage Unit Type:    Media Manager
Host Connection:      ebs6-bck
Number of Drives:     4
On Demand Only:       no
Max MPX/drive:        12
Density:              hcart2 - 1/2 Inch Cartridge 2
Robot Type/Number:    TLD / 1
Max Fragment Size:    1048575 MB

    Why is that when only 2 drives are assigned to this media server on this host?

    As I have said before, this will result in over-allocation of jobs for this STU.

    Can you confirm that SSO license key was added to both of these media servers? (Check output of get_license_key on both media servers).

    vmoprcmd tells us that NBU was probably not restarted on media server ebs7 after drives were configured as SSO?

    ESL025_Drive12           No      No                     No        hcart2
    ebs6-bck                   /dev/rmt/1cbn                        SCAN-TLD
    ebs7-bck                   /dev/rmt1.1                          RESTART

     

    NBU needs to be restarted on both media servers.

    Please correct the STU config on the master server and run 'bprdreq -rereadconfig'.

    If above steps were taken and drive 12 is still not used, try to down drive 11 on ebs6 and see if drive 12 gets used.

     

    Let us know the result.

19 Replies

Sort By
Show More
  • Iwan_Tamimi's avatar
    Iwan_Tamimi
    Level 6
    11 years ago

    Marrianne, 

    Thanks again for helping with the case.

    I attached all the information you requested.

    > Is reason for queued jobs still 'Drives are in use'?

    Yes. on storage unit ESL025_EBS6_MPX consist of 1 single drive and 1 SSO drive, when the single drive empty then the job continue to that drive.

     

  • Marianne's avatar
    Marianne
    Level 6
    11 years ago

    Thanks for the output.

    Firstly, we see that stu ESL025_EBS6_MPX  is configured for 4 drives:

    Label:                ESL025_EBS6_MPX
Storage Unit Type:    Media Manager
Host Connection:      ebs6-bck
Number of Drives:     4
On Demand Only:       no
Max MPX/drive:        12
Density:              hcart2 - 1/2 Inch Cartridge 2
Robot Type/Number:    TLD / 1
Max Fragment Size:    1048575 MB

    Why is that when only 2 drives are assigned to this media server on this host?

    As I have said before, this will result in over-allocation of jobs for this STU.

    Can you confirm that SSO license key was added to both of these media servers? (Check output of get_license_key on both media servers).

    vmoprcmd tells us that NBU was probably not restarted on media server ebs7 after drives were configured as SSO?

    ESL025_Drive12           No      No                     No        hcart2
    ebs6-bck                   /dev/rmt/1cbn                        SCAN-TLD
    ebs7-bck                   /dev/rmt1.1                          RESTART

     

    NBU needs to be restarted on both media servers.

    Please correct the STU config on the master server and run 'bprdreq -rereadconfig'.

    If above steps were taken and drive 12 is still not used, try to down drive 11 on ebs6 and see if drive 12 gets used.

     

    Let us know the result.

  • RadovanTuran's avatar
    RadovanTuran
    Level 4
    11 years ago

    Try to play with SCSI reservation on tape drives. it can be configured globally and the each drive is using "Default" or you can change it per tape drive. SCSI-2 reservation doesn't work everytime with SSO but SCSI-3 reservation is not supported by all devices.

    Play with both versions of config, check config of tape library. Many times firware upgrade change some values to defaults. Ask HP which SCSI reservation is supported by TL (SCSI-3 should be supported by new devices).

  • Iwan_Tamimi's avatar
    Iwan_Tamimi
    Level 6
    11 years ago

    Marrianne,

    > Why is that when only 2 drives are assigned to this media server on this host?

    Actually I wasn't sure why it was 4, now I undertand a bit,  I will change accordingly.

    >Can you confirm that SSO license key was added to both of these media servers?

    I think it should not be a problem, the configuration for both server have been in production for several month until the TapeLibrary upgrade problem. We didn't change any license config in both servers.

    I will restart the NBU and give the ressutl.

     

    Radovan,

    The drives connected through fiber through the Fiber switch. The system actauly has been run for sometimes.

     

    Regards, 

    iwan

  • Marianne's avatar
    Marianne
    Level 6
    11 years ago

    PLEASE double-check SSO licenses on both media servers.

    We saw a similar issue here about a week ago when drive control remained in RESTART.
    Issue was solved when SSO licenses were added to media servers.

    Media servers should have these 3 licenses:
    NBU Enterprise server (or Enterprise Client if SAN media server)
    Library based tape drives
    Shared Storage Option

  • Iwan_Tamimi's avatar
    Iwan_Tamimi
    Level 6
    11 years ago

    Marianne,

    Actually I also asked Symantec to help , and the support use the  "Configure Storage Devices" wizard on the Java GUI and after that seems ok. I didn't use the wizard when I configured it, I still don't know what I did wrong. Using the wizard , the support actually configure all the for drives to be SSO. After that I took out 1 drives from each media server (so each media server will get 3 drives which 2 of them are in SSO)then the problem come again.

    Then I see your finding about my STU config, I corrected to 3 then looks the backup running fine.

    Thank you very much.

    Iwan

  • Marianne's avatar
    Marianne
    Level 6
    11 years ago

    ALWAYS best to use the wizard with SSO config...

    Glad it's working now!