Tapes cannot be scanned back in when "Amount of blocks to read/write at a time" is set to > 1 (4045754)

Date: Jan 08
Product & Version: < 8
Module & Version: N/A
OS Version: All
Application version: N/A

Symptoms:
If NetVault is configured (Device Management -> Right-click on a drive -> Performance) to write and/or read more than the default one media block at a time, there can be issues scanning this tape back into the database once it has expired/been removed. The tape will lose its "Foreign" status, but not reveal any savesets for restore.

Reason:
When using n-blocks-at-a-time > 1 there is a problem with the way the indexes are written to tape. Normally the index is located immediately following the main datastream of the backup, aligned to the nearest block boundary - so there is a small amount of "padding" between the backup and the index to make the index itself start at the beginning of a block. This makes it possible to find the index again when scanning as NetVault knows exactly where to expect it.

Unfortunately, the n-blocks-at-a-time option inserts more padding - it aligns to a blocksize * blocks-at-a-time boundary. This stops scanning from working as it is expecting to find the index just a single block after the backup.

Solution:
The faults for this are NVG-1214, NVG-1305, NVG-1817 and NVG-1978. The first two faults have been addressed in NetVault 7.4.3, with the remaining two faults fixed in NetVault 8. Therefore, if you experience this issue you should upgrade to the latest version - the fix has been implemented in such a way that old tapes written with this setting can be scanned and restored from, so the data is safe.