HP StorageWorks MSL2024, MSL4048, and MSL8096 Tape Libraries and 1/8 G2 Autoloader Ultrium 1760 tape drive firmware release notes (AK378-96023, August 2009)

especially during tape recovery. Given that the library is able to disable/enable the primary ports,

then the restriction on configuring the external encryption control while volume is mounted has

been changed to the following: If no primary ports are enabled, then accept the command to

change settings. Furthermore, if there is no tape mounted in the drive and the ports are enabled,

then changing encryption control settings ought to be allowed too.

Drive hangs

Previously, under certain timing conditions, it was possible that a read operation could hang, due to

the firmware explicitly consuming an important signal. This issue is now resolved.

Write/read/spacing issues

• Previously, if the removal of the medium was prevented, pressing the eject button or an issuing an

unload command would not eject the tape, however the Logical Position would be wrongly reset

which could cause a later read failure. This issue is now resolved.

• Added improvements so that the performance log is more up-to-date and accurate

• In this code version, when spacing forwards on a tape which has more than 100 filemarks that

are unevenly spread over a half wrap, the drive will not try to work out where to start the search

from, but will now start searching from the current position.

• Previously, if the drive received a locate command while the drive was still performing a read

ahead operation, the locate command could return a good status without actually performing the

locate and a time-out could potentially occur. This issue is now resolved.

Interface-related issues

• Fixed an issue with Transport Layer Retry (TLR) handling in SAS drives that has been seen on DAT

drives as a precautionary measure.

• SAS drives specifically connected to HP’s new Proliant RAID smart array controllers (P212, P410,

P411, P712) were subjected to a large Transport Layer Retry (TLR) occurrence and as a result

performance was adversely affected. The performance is drastically improved by reducing the

frame credit to 1.

Automation issues

• Previously, customers using BackUp Exec software, could possibly experience “0x980a =

AD_NO_TASK_OBJECT” errors. This issue is now resolved.

• Previously, if a response timeout occurs during an Abort Task Management Function the data

connection to the library could hang. This has been fixed.

Universal cleaning cartridge

• Previously, in the very unlikely scenario that a Universal Cleaning Cartridge (UCC) had a stuck

leader pin, the drive would fail to eject it and would have to be power cycled. This issue is now

resolved.

• Previously, if a UCC was not initialized correctly it could be left with a thread count of 1. That

meant it would never be re-initialized and it would always attempt to clean in the same area of

tape and never increment the thread count. In simple terms, the end user could end up using this

UCC for an infinite number of cleans. This issue is now resolved.