Use of Scratch pool

[V4]

We have 3-tier NBU in our organization with 115 client server, 8 drive HP robotic library and 20hrs busy schedule.

NBU 7.0.1 (Master Server)

NBU 7.0 (Media Server)

We have 90 polices for all these servers and 40 volume pools according to application and types of backup.

I cam across scratch pool in netbackup, and heard that it can help me avoiding 96 error as i'm facing this issue randomely, despite of media being invetoried.

Can anyone brief about use of scratch pool and it's management.

Below is my requirement.

how netbackup searched media i.e. flow

1>assigned pool

2> netbackup

3>scratch .is this order correct.

does any notification or alert comes with media ID and policy name if tapes are used from scratch pool instead of assigned pool. This is our majority requirement as we are given media's by different app owner. and pools are created accordingly.

Cheers

Vaibhav

[Kiran_Bandi]

You can have only one Scratch pool configured with NetBackup.If a Scratch pool is configured, Media Manager moves volumes from that pool to other pools that do not have volumes available.Once those volumes become unassigned, those will be returned back to Scratch pool by Media Manager. If don't want those volumes to come back to Scratch pool you can use the following command to disbale that feature:

nbemmcmd -changesetting -return_unassigned_media_to_scratch_pool no -machinename <EMM host>

NBU serches for volumes in the assigned pool first and if not available it searches in Scratch pool for usable media.

Also You should not configure policies to use Scratch Volume Pool.

Regards.....

[Zahid_Haseeb]

I hope my below post will resolve all your concerns:

You also can configure a scratch pool from which NetBackup can transfer volumes

when a volume pool has no volumes available.

If a volume becomes full, NetBackup assigns

another volume for use if one is available in the pool or in a scratch pool.

When the number of partially full media is reached, NetBackup queues backup

jobs until media becomes available. If a media becomes full, NetBackup assigns

another media for use if one is available in the pool or in a scratch pool.

does any notification or alert comes with media ID and policy name if tapes are used from scratch pool instead of assigned pool. 

if a NetBackup policy or schedule specifies the scratch pool, all

requests for those volumes are denied.

You can find a proper documentation from the Netbackup Admin Guide for windows Page # 292 About scratch volume pools

[Andy_Welburn]

"This is our majority requirement as we are given media's by different app owner. and pools are created accordingly."

As mentioned in the first reply, NB can only have one scratch pool from which all backup jobs could call media if required. You would have absolutely no control over which media is called by which job which I think is what you require? If your "cilents" (customers) are determining which set(s) of media to use for which jobs then you'll just have to ensure that each pool has enough media to cater for your upcoming jobs.

[Marianne]

A scratch pool will certainly help to alleviate status 96's.

Extract from Admin Guide I:

About scratch volume pools
The scratch pool is an optional pool that contains the media that NetBackup can
allocate to other pools as needed. If you configure a scratch pool, NetBackup moves
volumes from that scratch pool to other pools that do not have volumes available.
Only one scratch pool is allowed.

When you create a scratch pool, you can call it anything you like, as long as you select the 'scratch pool' tick box. It just makes sense to call it something like Scratch.

It will require that you manually change all unassigned media to scratch. You will have to keep on doing it as and when images expire and media returns to unassigned. This is a once-off task for all media. After this, media will automatically be returned to scratch upon expiration. The following setting is enabled by default:

# nbemmcmd -listsettings -machinename <master> |grep -i scratch
RETURN_UNASSIGNED_MEDIA_TO_SCRATCH_POOL="yes"

You can also add a Barcode Rule for all new media to be put into Scratch Pool.

NetBackup will never search the NetBackup pool unless this pool name is explicitly specified in the policy.

how netbackup search media:

1. assigned pool
2. scratch

More info: http://www.symantec.com/docs/HOWTO32854

[V4]

hi Zahid,

If i use scratch pool, what should be configured in policy

for notification i would require media ID and policy name which had taken media from scratch pool.(If policy is configured without scratch pool)

This information helps us to track the usage of media. and for restoration purpose also.

if notification is not possible then it becomes difficult for admin to track media usage.everytime admin would need to open job propertise and from details pane he would need check for media used. this manual intervention would be time consuming along with day to day regular activities.

Cheers

Vaibhav

[rizwan84tx]

You may set the scratch pool as default volume pool on inventorying robot, by below bar code example rules.

Barcode tag: <DEFAULT>

Media type: hcart3

pool name:  Scratch Volume Pool

All backups will be using the media from scratch pool and hence at the end of backup you may run a report and track the media usage.

[Zahid_Haseeb]

If i use scratch pool, what should be configured in policy

You dont need to configure any thing in the policy. When your Media will full the Netbackup will automatically go to search the Scratch Pool to find the Media

for notification i would require media ID and policy name which had taken media from scratch pool.

If Netbackup writes any Error/Warning/Info( which i am not confirm) in the Event Viewer when the media being assigned from scratch pool in case of existing Media Full then i think i may possible. See the below Blog it may help you i hope.

https://www-secure.symantec.com/connect/blogs/howto-check-netbackup-tape-drive-status-and-generate-e...

OR

try to use and search Opscenter admin guide

[V4]

Hi Zahid,

"in OPs Center we need to execute query" this was statement by Sym Support. Though assigned engg told that using query we can get this details but we need to identify which table does it points to.

i thought this would be available in OPs Center but didn't got any reports which gave give this details.

For scratch pool i believe your point of view comes after my existing media is full.

but i'm asking at beginning of backup will netbackup pick any media from scratch pool.

[Marianne]

"at beginning of backup will netbackup pick any media from scratch pool."

Only when there is no suitable media in the required pool to write to.

See Admin Guide I:

"NetBackup moves volumes from the scratch pool to other pools that do not have volumes available."

Also see http://www.symantec.com/docs/HOWTO32854

[watsons]

Hi Vaibhav,

By now you probably realize how effective a Scratch pool can avoid error 96, but I think your dilemma to implement is that how you are going to track the media usage for each business unit, because you simply don't own the media. Your business unit purchase them for you, based on reported usage.

From my experience, when we talk to different business unit we don't give them media usage but data usage. It makes more sense and is perfect for us to use Scratch pool.

We ask them how much data they want us to backup, and we calculate the min. amount of media they need to purchase. In this case, we will have a bunch of media from each business unit and it's up to us to manage & use (put all media into Scratch pool). Business unit does not really need to know what tapeID they use, but they want to know how many media they need to or had purchased - about $$. That's all.

That will make life easier, for us and for them :)

My 2 cents.

[V4]

i agree with your view watsons. we are not gonna make it transparent to BU's. they are just bothered about media count. But our concern was to track them for retore purpose.

For same i raised query if media ID can be populated with scratch pool. so it becomes easier for us to update them in our daily sheets. If OPs Center can provide(still couldn't figured out which reports does it shows.) then my all issues are resolved.. even this may be useful for auditing purpose... (this was demanded by BU's ) can't get it how this can be useful in terms of audit. It may be which i am unaware.

[Kiran_Bandi]

For restoring NBU will take care about which media is required for restoring. If that media is present in the library, NBU will use it for restoring. Otherwise it prompts you to insert it into the library to continue. 

Regards....

[watsons]

I have not much experience with OpsCenter or the older Backup Reporter, what we used to do with keeping track of media usage is by writing script. 

Use of bpmedialist or bpimagelist can tell us what mediaID are being used for daily or weekly backup etc.

[V4]

it sounds Scripts can solve it.. watson could you please post test solution of the script ..

[watsons · Accepted solution]

Sorry Vaibhav, I can't share with you the entire script, but I can provide you what commands we used in it:

# bpimagelist -media -idonly -hoursago 144 -st full    ==> Retrieve all full backup mediaID used in last 144 hours (6 days).

For daily report, you can use -hoursago 24 and exclude -st, that would give you all full/incr/cinc. Depend on your requirement.

In my script, as we want to offiste the full backup media, we use:

# vmquery  -l -m <mediaID>    

for those mediaID we found in first command output, that would give us the slot number in the library and what volume pool they belong.

Hope this helps.

[V4]

Thanx watson,

Would be trying in my test lab as well in prod. It seems this scripts would resolve my media tracking issue.