HP Tru64 UNIX and TruCluster Server Version 5.1B-5 Patch Summary and Release Notes (March 2009)

Fixes a klog inconsistency during node boot in which it is possible that no nodes
in the cluster respond to a klog sync request and the booting node simply clears
the disk group klog.
Decreases CLSM boot times in large cluster configurations.
Fixes problems in the vol_new_disk ioctl when errors are encountered when
loading a disk in the kernel.
Fixes the cause of a panic in LSM voldrl_commit_delete that occurs during a
transaction abort.
Corrects the cause of a panic in LSM voldrl_commit_delete seen during a
transaction abort.
Adds performance enhancements to help reduce LSM cluster-wide plex-detach
processing time.
Fixes a problem whereby LSM does not properly removing plex detach entries
from its kernel change log in a TruCluster environment.
Fixes an issue with kernel plex detach by attempting to further reduce the number
of redundant detach events sent from the kernel to vold.
Fixes a boot problem with the LSM I/O performance change when the root file
system is an LSM volume.
Adds a configurable fastfail option to the LSM driver for mirrored volume I/O.
Modifies conditions to allow multiple retries to mount the root file system and
provide the ability to adjust the retry period.
Corrects a locking issue in KCL local lock.
Fixes an "assert wait" panic caused by automatic LSM error tracing when DRD
uses LWC (light weight context) switching and many I/O and plex detaches occur.
Fixes a problem in which I/O requests hang when multiple LSM kernel-initiated
plex detaches are processed simultaneously by different cluster members across
different diskgroups.
Patch 27052.00
OSFLSMX11540
Prevents a process hang occurring when large I/O's are performed on multiple
drl-enabled LSM volumes simultaneously on Multi-CPU systems.
Causes zero length I/O requests to be flagged with the proper error message to
the user rather than causing a panic.
Fixes a “clsm_config" dlm_lock deadlock issue.
180 Tru64 UNIX Patches