Highlighted

NetBackup 8.2 - SLP Backlog question

NetBackup 8.2 on a Veritas 5230 Appliance running 3.2 plus patches.

I'm looking to see if there is a way to clear a backlog of SLPs that are supposed to be doing to an S3 bucket or to tape?

I have a list of SLPs on a single master server with multiple media servers and Clound Catalyst servers that has a backlog of about 1 PB of data. 

Is there a way to list what's waiting to be processed and then a way to clear them out? nbstlutil shows about 6,421 copies waiting to be processed?

Tags (1)
8 Replies
Highlighted

Re: NetBackup 8.2 - SLP Backlog question

If by "clear" you mean "cancel", then yes, you can do that but note that if it's cancelled it cancelled. You'll have to manually duplicate those images if you want to send them somewhere else.

You can use to see the incomplete images

nbstlutil list -image_incomplete

You can cancel one you need to with

nbstlutil cancel [-lifecycle name |

Highlighted

Re: NetBackup 8.2 - SLP Backlog question

@DPeaco@RiaanBadenhorst has already suggested options to look into the SLP backlogs.

I'm adding to it, please check it may help you.

nbstlutil report => SLP Status
nbstlutil list -image_incomplete -lifecycle name_of_lifecycle => Incomplete or pending SLP

There is chance that your SLP might be inactive so active based on what has to be done in you environment.
nbstlutil active -lifecycle name_of_lifecycle

Highlighted

Re: NetBackup 8.2 - SLP Backlog question

nbstlutil list -image_incompete

Yielded a list of 57,255 images that are waiting to be completed?

How do I target a specific image in the "queue" to cancel it?

I'm asking because several of these are already expired based on Policy/SLP set retention. I just need to clear the backlog of already expired SLPs.  

Highlighted

Re: NetBackup 8.2 - SLP Backlog question

IMO, with such a large backlog of so many images to consider, and a desire to filter by multiple criteria (age, client, destination, policy, type) and decide which to cancel versus which to continue re-attempting... then...IMO (and it is just an opinion) then your only option is scripting.

If you have a Windows based Master Server, or a Windows based Media Server with server rights over the Master... then I might be able to help you... as I have a natty little batch/cmd plus VBScript combo which can filter and report and generate the required "cancel" commands for you... but... BUT... BUT BUT... it is quite a rough script and not for the faint hearted... so... if you have the skills and the patience then I might be able to help you with a copy of them.

It comes in two parts.  1) a batch/cmd script to collect various text/list files from NetBackup    2) a VBScript to read the lists, munge the meta-data, and then spit out        2a) multiple reports/tables via different sorted groupings       2b) a commented out batch/cmd script to perform SLP image cancels.

So, these scripts are utterly safe to run (in the first stage) as long as your environment can handle being listed... as the scripts will change nothing at all, except consume perhaps a few MB of disk space for text/list files.

Highlighted

Re: NetBackup 8.2 - SLP Backlog question

@DPeaco 

1. Extract a list of images in Excel which are still there beyond expiry date from Netbackup Catalog Report, but cant go away due to pending slp

2. In next column in front of each slp , type nbstlutil cancel  -backupid

3. In Next Column  do a concatenate of 2 columns

4. This will give you a command in column 3 to cancel that specific backup ID SLP

a row in your excel should be similar to this

  

    A                                B                                   C

clientname_11252252566   ./nbstlutil cancel -backupid      ./nbstlutil cance -backupid clientname_11252252566

copy all entries in column 3 to a file, rename it as a script and execute it on server.

Its the simplest way to batch clean the SLPs

Highlighted

Re: NetBackup 8.2 - SLP Backlog question

sdo

Thanks for your reply. We are a Unix / Linux shop but we do have windows backup "clients". I am not a stranger to scripting via KORN shell, BASH, and even some Python script work. I just need to find out how to get the info I need and then what I can do with it to clean up. :-) 

Highlighted

Re: NetBackup 8.2 - SLP Backlog question

Hi @DPeaco 

Here is a Perl script I have used successfully in the past to cancel SLP processing for images that should have already expired. Internally it uses a combination of bpretlevel, nbstl & nbstlutil etc to build up a picture of what should have expired but hasn't due to unprocessed SLP jobs. It creates 3 files - the main one being the command file that contains the list of images to cacnel/expire. So running the script by itself will not do anything to the environment - you have to run the command file created to do the work post running this script). 

Use it at you peril - so after running my script, sanity check the created command file to make sure some of the images that it will cancel/expire are valid.

Another question you need to have answered is why the backlog is so huge - are you trying to do too much, are your network links too small, or was this some historical anomoly where you had to disable duplication/replication for a time and now need to catch up.

Good luck

Highlighted

Re: NetBackup 8.2 - SLP Backlog question

davidmoline,

Thank you! 

To answer your question as to why the backlog....we've had Cloud Catalyst issues for the past 2 years, sometimes working and sometimes not. There are other things that happened like having to rebuild CC servers and such due to corruption. We are a large shop. We do over 10,000 backups per day, with an annual total of over 120 PB of data. A mix of dedicated solutions as well as large NBU shared solutions. Right now, I'm trying to sort through a 1 PB backlog of SLPs to help reduce the loads on the masters. Things like this happen easily...especially when you continue to grow data and someone up the chain doesn't allow to spend money to upgrade data protection infrastructure. We do have things "in the works" to make things better, but we're still having to do a lot of "back end" work to try and keep it all together.