cancel
Showing results for 
Search instead for 
Did you mean: 

Only 3 of 4 drives used in NetBackup 7.1.

Willie_SA
Level 6
Partner Accredited

Only 3 of 4 drives used in NetBackup 7.1.

NetBackup                          7.1

Master/Media server    UNIX Solaris 10

 

I have 4 x LTO5  Drives in SAN configured and online and working.

BUT NetBackup only use 3 drives at a time. Not always the same 3.

 

I did check:

License is OK.

Increase and decrease number of policies

Limit jobs per policy = 25

Allow multiple data streams = Ticked.

Media Multiplexing = 4

 

What else can I check?

1 ACCEPTED SOLUTION

Accepted Solutions

Nicolai
Moderator
Moderator
Partner    VIP   

When no job are running execute command: nbrbutil -dump

No allocation shoud be shown. If there is allocation run  nbrbutil -resetMediaServer <mediaserver> or

/usr/openv/netbackup/bin/admincmd/nbrbutil -resetAll

nbrbutil doc link:

http://www.symantec.com/business/support/index?page=content&id=HOWTO43779

View solution in original post

17 REPLIES 17

Mark_Solutions
Level 6
Partner Accredited Certified

Four things come to mind:

1. Multiplex level set on all schedules is 4

2. Open up the storage unit for the library to ensure that it says it has 4 drives

3. Maximum jobs per client (master host properties)

4. That they are all of the same and correct density (HCART or HCART2 or HCART3 - but not a mix)

Hope this helps

Nicolai
Moderator
Moderator
Partner    VIP   

GUI -> Storage -> Storage Units -> Open the Storage unit in Qustion.

Check the value of "Maximum Concurrent Write drives". Must be 4 in order for all of yor drives to work at the same time. 

Willie_SA
Level 6
Partner Accredited

Double checked - > Storage Unit -> Mac Concurrent Write Drives = 4

Was =4 all the time.

Mark_Solutions
Level 6
Partner Accredited Certified

As well as the items I mentioned above I have thought of one more thing ...

A tape could have been left in the drive

When nothing is running use robtest to see if a tape is in the drive, if there is use robtest to move it back to its slot

If you are not familiar:

run it from volmgr\bin\robtest

then enter:

s d

This shows the drives - if there is a tape in one it should show it and which slot it came from - if it doesnt show the slot then use s s to show the slots and make a note of an empty one

Then use:

m d1 s22

which in this example would move it from drive 1 to slot 22

The drive should then be available again

mph999
Level 6
Employee Accredited

OK, run a test ...

Turn off multiplexing (on the storage unit) - but leave on multiple streams (in the policy).

Run a policy that will create at least x4 streams 

Does it now use all x4 drives.

 

I think you have done, but make sure media multiplexing is set to 4 on all policies.

We do not known exaclty what you are doing, but, are you running enough streams to use x4 drives - in theort the 'first' 12 streams will go to x3 drives, if there are not enough 'jobs' running , the 4th will not be used.

I am sure you have checked this last point, but I'd better ask ...

 

Martin

Willie_SA
Level 6
Partner Accredited

1)    Multiplex level is = 4  (Media multiplexing)

2)    STU = does not show nr of drives. - Am I missing something?

3)    Max Jobs per client = 25

4)    All Drive Type = HCART2

Willie_SA
Level 6
Partner Accredited

When jobs run , in Activity Monitor:

12 jobs will run and any more than 12 will QUEUE.

Can not get more than 12 to run.

phil_scarfo
Level 3
Partner Employee Accredited Certified

All the above suggestions are good.

If those don't seem to provide a resolution, check to make sure that tapealert for cleaning did not get set and clear that.

also, make use of nbrbutil command to check that there isnt a device allocation hanging out there.

good luck.

Willie_SA
Level 6
Partner Accredited

I can only so ROBTEST tomorrow

have not done it before,

BUT did physically check on robot and drives - and there were NO tapes on the

NON USED drive.

Will double check tomorrow

The_Director
Level 5
Certified

Are these all set?

 

Master Server Properties > Global Attributes > Maximum Jobs per client : >12

Storage Unit > Maximum Concurrent Write Drives : =4

Storage Unit > Enable Mulitplexing : checked

Storage Unit > Maximum Streams per Drive : >4

Policy > Attributes > Limit Jobs Per Policy : >12

Policy > Schedule > Attributes > Media Multiplexing >4

 

 

Also make sure you have 4 available tapes and that the 4th is not frozen, suspended, or full.

Nicolai
Moderator
Moderator
Partner    VIP   

When no job are running execute command: nbrbutil -dump

No allocation shoud be shown. If there is allocation run  nbrbutil -resetMediaServer <mediaserver> or

/usr/openv/netbackup/bin/admincmd/nbrbutil -resetAll

nbrbutil doc link:

http://www.symantec.com/business/support/index?page=content&id=HOWTO43779

Willie_SA
Level 6
Partner Accredited

Result from Robtest

 

No carts on drives.

 

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

Robot selected: TLD(0)   robotic path = /dev/sg/c0tw500104f000aff128l0

Invoking robotic test utility:
/usr/openv/volmgr/bin/tldtest -rn 0 -r /dev/sg/c0tw500104f000aff128l0

Opening /dev/sg/c0tw500104f000aff128l0
MODE_SENSE complete
Enter tld commands (? returns help information)
s d
drive 1 (addr 1000) access = 1 Contains Cartridge = no
drive 2 (addr 1001) access = 1 Contains Cartridge = no
drive 3 (addr 1002) access = 1 Contains Cartridge = no
drive 4 (addr 1003) access = 1 Contains Cartridge = no
READ_ELEMENT_STATUS complete

Willie_SA
Level 6
Partner Accredited

Result from nbr butil:

Does thois mean that there is allocation stuck for media ID R00024?

R00024 shows in GUI => Medai as 0 times mounted?

 

Allocation Requests
(AllocationRequestSeq )


Allocations
(AllocationSeq )

MDS allocations in EMM:

        MdsAllocation: allocationKey=831 jobType=16 mediaKey=4000093 mediaId=R00024 driveKey=0 driveName= drivePath= stuName= masterServerName=cprvnmsnetb mediaServerName=cprvnmsnetb ndmpTapeServerName= diskVolumeKey=0 mountKey=0 linkKey=0 fatPipeKey=0 scsiResType=0 serverStateFlags=0

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Not sure that allocation for R00024 is causing the problem as no drive is associated with it.

If no jobs are currently running, clear the allocation:

nbrbutil -releaseMDS 831

Silly question, but are all 4 drives UP and configured correctly?

vmoprcmd -d

tpconfig -l

(these commands live in <install-path>\veritas\volmgr\bin)

The_Director
Level 5
Certified

If all the settings in the post are set and you have specifically 1 tape that is not mounting then I am going to assume that the issue is with that tape. Do you have any other tapes you can assign to the pool you are writitng to or even assign an extra tape to the scratch pool and run the job to see if it grabs a new tape for that 4th job.

RUN /usr/openv/volmgr/bin/vmquery -m R00024 to see that status of the tape

Willie_SA
Level 6
Partner Accredited

The problem is not with one tape, it is with one tape DRIVE.

4 drives configured and UP, but only 3 drives are being used at one time.

It is not always the same drive NOT being used.

Tonight drive 1, 2 and 4 OR tomorrow night 2,3 and 4 ...............

Willie_SA
Level 6
Partner Accredited

After clearing with nbrbutil there were no more allocation.

Tested tonight and all 4 drives are now being used.

Problem fixed.

Thank you