During a hot catalog recovery, image cleanup may run and will 'cleanup' newly restored .f files.
At a VERBOSE level of 5, the bpdbm log will show image cleanup finishing and NOT deleting any files:
17:38:35.483  <4> db_error_add_to_file: deleted 0 expired records, compressed 0, tir removed 0, deleted 0 expired copies
17:38:35.489  <4> ImageDelete::~ImageDelete: deleted 0 expired records, compressed 0, tir removed 0, deleted 0 expired copies
Review of an ls -lRa or dir /S of the images directory will show that the images that were restored are no longer there.
Upon further review of the bpdbm log it can be noted that the cleanup process actually deleted numerous "unused" <ctime> subdirectories:
17:38:31.436  <2> ImageDelete::delete_unneeded_timestamp_dirs: Found 55, deleted 54
17:38:31.450  <2> ImageDelete::delete_unused_reserved_images: Deleting unused reserved images ...
17:38:31.470  <2> ImageDelete::delete_unused_reserved_images: Successfully expired unused reserved images
17:38:31.470  <2> ImageDelete::delete_unused_ghost_images: Deleting unused ghost images ...
This issue is caused by a defect that allows image cleanup to run while a hot catalog recovery is in process.
The formal resolution for this issue (Etrack 3286467) is included in the following releases:
- NetBackup 7.6 GA
- NetBackup 7.5 maintenance release 7 (126.96.36.199)
These releases are now available - more information and download information is available at the following link:
The touchfile /usr/openv/netbackup/bin/NOexpire (UNIX) or <install_path>NetBackup\bin\NOexpire (Windows) can be put in place to prevent image expiration.