cancel
Showing results for 
Search instead for 
Did you mean: 

Failover storage unit group

nbujobber
Level 3

We have a seem to be having a problem with a failover storage unit group.As I understand it,it should only failover if there is a problem.

Assoon as this storage unit reaches its configured number of jobs,the following will not queue but go the configured failover storage unit.

Any ideas ???????

 

12 REPLIES 12

RamNagalla
Moderator
Moderator
Partner    VIP    Certified

 

Failover

If the Failover option is selected, when a job must queue for a storage unit, the job queues rather than try another storage unit in the group.

 

A queue can form for a storage unit if the storage unit is unavailable.

The following are some reasons why a storage unit can be considered unavailable:

  • The storage unit is busy.

  • The storage unit is down.

  • The storage unit is out of media.

  • The storage unit has no available space.

  • The storage unit has reached the Maximum concurrent jobs setting.

 

so your jobs should be in Queue.

 

does your storage unit is local storage unit? both media server and cliets are same?

does the same storage unit is configured in any other group?

 

nbujobber
Level 3

Storage units are local.Both media servers are AIX6.storage unit group configured with 2 storage units as failover

RamNagalla
Moderator
Moderator
Partner    VIP    Certified

 

Exception to the storage unit selection criteria on Windows

The only exception to the storage unit selection criteria order is in the case of a client that is also a media server with locally connected storage units. The locally available storage units take precedence over the defined sequence of storage units in the group.

http://www.symantec.com/business/support/index?page=content&id=HOWTO34501#v41275477

 

I am assuming it would be same for Unix servers also.

 

so the jobs that are using the 2nd storage unit are local to the jobs?

nbujobber
Level 3

We only have storage unit groups.This group has 2 storage units configured as failover.one on Server A, The other on Server B ,but the failover takes place as soon as the limit of the first storage unit is reached,instead of queueing

RamNagalla
Moderator
Moderator
Partner    VIP    Certified

could you provide the output of below commands

 bpstulist -label <1st stuname> -U

bpstulist -label <2nd stuname> -U

bpstulist -group  <STu Group>

nbujobber
Level 3

bpstulist -group I6866610_ACS_LANA_LTO4 -U
I6866610_ACS_LANA_LTO4 3 I6866610_ACS_LANA_LTO4_1200 I6880010_ACS_LANA_LTO4_1200
 

bpstulist -label I6866610_ACS_LANA_LTO4_1200 -U

Label:                I6866610_ACS_LANA_LTO4_1200
Storage Unit Type:    Media Manager
Host Connection:      i6866610bu.sbb.ch
Number of Drives:     4
On Demand Only:       yes
Max MPX/drive:        24
Density:              hcart - 1/2 Inch Cartridge
Robot Type/Number:    ACS / 1
Max Fragment Size:    15000 MB

bpstulist -label I6880010_ACS_LANA_LTO4_1200 -U

Label:                I6880010_ACS_LANA_LTO4_1200
Storage Unit Type:    Media Manager
Host Connection:      i6880010bu.sbb.ch
Number of Drives:     4
On Demand Only:       yes
Max MPX/drive:        24
Density:              hcart - 1/2 Inch Cartridge
Robot Type/Number:    ACS / 1
Max Fragment Size:    15000 MB

RamNagalla
Moderator
Moderator
Partner    VIP    Certified

 

nbujobber,

i am not seeing anything odd on above config,

 

1)how many clients you are trying for this test?

you have 2 media server 

i6866610bu.sbb.ch

i6880010bu.sbb.ch

so when Failover is taking place for I6866610_ACS_LANA_LTO4_1200 to I6880010_ACS_LANA_LTO4_1200, what is the client name that is using that 2nd stoarge unit?

does it the 2nd Media server as a client? or some other Client?

the reason I am asking this is , I have read the below statement in tech note

The only exception to the storage unit selection criteria order is in the case of a client that is also a media server with locally connected storage units. The locally available storage units take precedence over the defined sequence of storage units in the group.

http://www.symantec.com/business/support/index?page=content&id=HOWTO34501#v41275477

 

if that is the 2 nd media server as a client could you try the backup with other clients by excluding the 2nd media server for backups?

 

 

 

 

nbujobber
Level 3

No clients are media servers,we tried adding the second storage unit again last night,and the same thing happend.at some stage the backups start running on the second storage unit as well.as of this point on both are active

RamNagalla
Moderator
Moderator
Partner    VIP    Certified

its Really Odd behiaver, what is your Netbackup Versions and OS,

Probably you would need to get the  Symantec support to check if they have any Known bugs with this.

nbujobber
Level 3

OS .AIX 6.1 Netback 7.1

nbujobber
Level 3

problem still exists

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

I agree with Nagalla - please log a Support call with Symantec.

We can see from bpstulist output that selection method is indeed a "3":

bpstulist -group I6866610_ACS_LANA_LTO4 -U
I6866610_ACS_LANA_LTO4 3 I6866610_ACS_LANA_LTO4_1200 I6880010_ACS_LANA_LTO4_1200

Extract from Commands manual:

 

Prioritized = 1 (default)
Least Recently Selected = 2
Failover = 3
Load Balance = 4
 
Option 1: Prioritized, selects the first storage unit in the list until the unit is
down, is full, or its max-concurrent-jobs setting is reached. Then the next
storage unit in the list is examined and so on until an available one is found.
Option 2: Least Recently Selected, selects the least-recently selected storage
unit.
Option 3: Failover is the same as Prioritized except MDS queues a job to wait
for the first storage unit if the max-concurrent-jobs is reached. MDS moves
on to the next storage unit in the list only if the first unit is down or full.
Option 4: Load Balance. If the user selects this option and has installed the
capacity management License key, Media Device Selection (MDS) balances
the job load.
 
You seem to be experiencing Option 1 behaviour.
 
I cannot see anything wrong with your config, so best to get a call logged with Symantec.
Hopefully you are on NBU 7.1.0.4 and not unpatched version?