NetWorker: How to delete Multiple or Individual SSIDs

About the NetWorker saveset lifecycle

• When a saveset is first created, it is recorded in the database with its configured retention date (clretent, in mminfo). This is the date after which the Administrator has elected to have the saveset automatically removed by NetWorker.
• When a saveset is cloned, that clone also receives a record with its own retention date. The mminfo value ssretent lists the last expiring copy of the saveset in the entire datazone.
• When nsrim runs (either manually or daily via Server Protection > Server backup > Expiry Action), all savesets are checked whether they are past their individual retention (clretent).
• If a saveset is past its retention date, but has dependent savesets which have still not passed theirs (e.g. a full backup, and its dependent incrementals), that saveset is not marked as expired (as it is still needed by its dependents).
• If no dependencies precluding expiry are found, the saveset is marked as Expired, and for disk volumes, NetWorker will immediately attempt to delete the data files corresponding the saveset record.
  • If deletion is successful: the record is also deleted.
  • If deletion is attempted, but fails for any reason: The record is marked with both an Expired flag (E) and Suspect flag (s) in the saveset instance's clflags field. The next time nsrim / Expiry runs, deletion will be reattempted, but even if it fails, the record is deleted anyway.
  • Because NetWorker will delete saveset structures for a volume which have no record in the media database anyway, deleting the saveset record skips the Expiry and retention calculation, and instead simply deletes files from the disk during the final phase of the Expiry action / nsrim run (space recovery).
• Once the files are deleted (by the Expiry or space recovery process), for a standard disk device, the space becomes available immediately. For a Data Domain-type device, the files are deleted, but the space will not be reclaimed until Data Domain filesystem cleaning has been completed. The same is true for a DD Cloud Tier-type device; after NetWorker deletes the record, followed by the saveset files, the space is still not reclaimed until Cloud cleaning has run.

Best Practices before Proceeding

• If the NetWorker server is a Virtual Machine, create a snapshot of the NetWorker server.
• Disable cleaning on the Data Domain.
• Perform a NetWorker Server Protection backup. Once completed collect the latest bootstrap backup details.

mminfo -B

• If file system space permits, create a copy of the following NetWorker database folders (res, mm, index). 
  • These folders are in the /nsr directory of the NetWorker install path.
  • The (mm) can be exported to an .xdr copy: How To Export the NetWorker Media Database Using nsrmmdbasm
• Rename the nsrim binary on the NetWorker server. This prevents accidental deletion. The process outlined in the KB involves renaming the binary back to default once ready to proceed with data deletion.
  • Windows: C:\Program Files\EMC NetWorker\nsr\bin\nsrim.exe
  • Linux: /usr/sbin/nsrim

How to delete Individual or Multiple SSIDs?

Prerequisite: This process can only be used for deleting savesets on Advanced File Type Devices (AFTD), File Type Devices (FTD), and Data Domain (ddboost) devices. Individual savesets cannot be deleted from a tape volume, even if that tape volume is on a Virtual Tape Library (VTL). The entire volume must be relabeled in order to reclaim space.

Examples of how to list SSIDs:

(quotes are needed around the query when using spaces or special characters):

• To display all savesets for a single media pool:

SYNTAX:

mminfo -avot -q pool=<mediapool>

EXAMPLE:

mminfo -avot -q pool=LinuxPool

• To display all savesets for a specific Client and two separate Volumes:

SYNTAX:

mminfo -avot -q client=<client_hostname>,volume=<vol1_name>,volume=<vol2_name>

EXAMPLE:

mminfo -avot -q client=linuxhost1,volume=LinuxPool.001,volume=LinuxPool.002

• To display all nonbrowsable Data Domain and adv_file savesets

EXAMPLE:

mminfo -avot -q  type=Data Domain,type=adv_file,recoverable 

• To display all savesets created over 30 days ago:

EXAMPLE:

mminfo -avot -q  sscreate<=30 days ago 

•  Display all savesets for a specific Group and Date range (quotes are needed when using special characters):

SYNTAX:

mminfo -avot -q  group=<group_name>,sscreate>=<start_date>,sscreate<=<end_date> 

EXAMPLE:

mminfo -avot -q  group=WinBackup,sscreate>=1/1/2013,sscreate<=2/1/2013 

The NetWorker Command Reference Guide also provides details on the mminfo command.

For INDIVIDUAL SSID deletion:

If a saveset ID without the clone ID is specified, it deletes all instances of that saveset, including the clones:    

nsrmm -o recyclable -S <ssid> | nsrmm -o recyclable -S <ssid/cloneid>

 nsrim -X 

The deletions are recorded in the daemon.raw log, and also reported in the NMC user interface log under the Maintenance tab.
If deletions against a specific volume are not occurring, verify that the operation is running:     

nsrstage -C -V <volume name>

For MULTIPLE SSID deletion:

Deleting multiple savesets using batch or shell script commands. 

1. First, append the mminfo output to only report the said (/cloneid):

SYNTAX: 

<mminfo command> -r ssid,cloneid -xc/

EXAMPLE:

mminfo -avot -q client=linuxhost1 -r ssid,cloneid -xc/

2. Next, redirect the output to a text file to be ran in a batch command:

SYNTAX:

mminfo command -r ssid,cloneid -xc/> > ssid.txt

EXAMPLE:

mminfo -avot -q client=linuxhost1 -r ssid,cloneid -xc/ > ssid.txt

3. Once the savesets list is created, confirm it is formatted correctly, with a single ssid/cloneid pair per line, ensuring there are no spaces which may separate the two. Since NetWorker 19.7 you can issue delete commands using an input file, such as the one just created:

nsrmm -dy -i ssid.txt

However, it is safer and preferred to mark each saveset desired for deletion as Expired. This requires a loop command to process each individually, and will take longer, but is safer as NetWorker will be able to better calculate dependencies when running Expiry Action each day:

WINDOWS: 

for /f %%A in (ssid.txt) do @nsrmm -o recyclable -y -S %%A

LINUX:

while read SID; do nsrmm -o recyclable -y -S ${SID}; echo ${SID}; done < ./ssid.lst

4. Manually start nsrim to perform the deletion process in NetWorker:

nsrim -X 

5. The deletions are recorded in the daemon.raw log, and also reported in the NMC user interface log under the Maintenance tab. If deletions are not occurring against a specific volume, verify that the operation is running:   

nsrstage -C -V <volume_name>

NetApp
For NetApp AFTD devices, the space is not immediately reclaimed, depending on a couple of factors:
If snapshots are taken of the CIFS/NFS file system where the adv_file device resides, they must expire before any data becomes available for cleaning. If the filer is under heavy load, deletions may take an extended period because space reclamation is a low-priority background operation.
 
Data Domain
For Data Domain DD Boost devices, freeing up space requires the Data Domain Cleaning operation to run.
By default, Data Domain only operates disk cleaning once a week on Tuesday. The cleaning cycle can be initiated from the Data Domain Enterprise Manager user interface (Data Management > File System > Start Cleaning).

See Also:

• NetWorker: Data Domain capacity increasing to 100% full - NetWorker side Troubleshooting (user log in to Dell Support required to view article)