Forum Discussion

Arun_K's avatar
Arun_K
Level 6
12 years ago

image clean up

Image clean up job expires the image  from the image database .   But what about the data which exist on disk? I have my below data on st. unit under /vol/data mount point.   -rw------- 1 r...
  • mph999's avatar
    12 years ago

    This is the best explanation, from Bob Insko, who is one of the best people in Symantec for anything to do with catalogs  (from the link Marianne provided)

    You will not get a better explanation than this :

    ****************************************************************************************************************

    Keep in mind that expiration of Disk images is a two part process.

     

    bpexpdate will detect that the image has past it's expiration time (or can be run manually).  If bpexpdate encounters a Disk Image, it will record the fragments of that image that exist on disk in EMM and they will be marked "To be deleted".  bpexpdate will then delete the header and .f file from the image directory.

    Later, nbdelete will run.  It will examine the appropriate EMM tables looking for fragments to actually delete from disk.  The appropriate media server will be contacted and bpdm on that media server will actually delete the fragment from disk.  If the media server cannot be contacted, or if the media server can't access the Disk STU, then the entry will remain in the EMM table so that an attempt to delete it can be completed on the next run of nbdelete. 

    Both of these jobs show up in the Activity Monitor as an Image Cleanup job.  Looking at the Job Details will show if a particular Image Cleanup job is a bpexpdate job, an nbdelete job, or both.

    nbdelete does not really "detect" expired images.  It only deletes fragments on images that were previously deleted by bpexpdate.

    One VERY DANGEROUS flag on nbdelete is the -force flag.  What that flag does is make THIS run of nbdelete attempt to delete the disk fragment from disk, but if that attempt fails because the media server is not available, or the media server can't access the disk STU (the disk STU is not currently mounted, for example), then even though the ATTEMPT to delete the disk fragment from disk FAILED, the entry in the EMM table will be deleted.  This means that if that media server comes back, or the disk STU is again available, no further attempts will be made by NetBackup to delete that fragment from disk by any automated means.  This would "orphan" (i.e.. abandon) that disk fragment on the disk STU where those files are just "taking up space" forever unless some other action is taken to delete them.