HP 3PAR StoreServ 7450 4-node HP 3PAR OS 3.1.2 Release Notes - Page 14

In previous versions of HP 3PAR OS, 3PAR WWNs which were, Default NL CPGs will use FC drives

Get HP 3PAR StoreServ 7450 4-node PDF manuals and user guides View all HP 3PAR StoreServ 7450 4-node manuals
Add to My Manuals
Save this manual to your list of manuals

Page 14 highlights

Issue ID Item Description 70797 VV state returns to "normal" if a tune In previous HP 3PAR releases, when a VV tune is cancelled the is cancelled displayed VV state remained at ""tuning"" indicating that the tune must be completed (re-run) or rolled back in order to return the VV state to normal. This behavior changes in 3.1.2 as the VV state returns to ""normal immediately on task cancellation. 71127 sysmgr panic - too many chunklet regions specified Avoid potential sysmgr failures by disabling reclaim for a volume if the total number of vlmaps exceed a certain threshold. 71684 Node state properly indicated after When a node is rebooted before the system is fully started, we will early node reboot now properly report its status in shownode. 72505 Potential for interaction between LD node balancing and importvv operations In rare cases a system that was trying and failing to rebalance LD node ownership while using Peer Motion to import volumes from another array could get stuck while doing the final VV block to finish the import. This was a volume block locking issue and has been resolved to prevent this negative interaction. 72983 Meta Data Inconsistency in bitmaps A type of meta data inconsistency when discovered caused the node can cause node to panic to panic rather than alert the user via the raising of the Internal Consistency Event. 73067 serivcehost list performance improved The servicehost list command will now have greatly reduced memory consumption and now focuses on the important information. 73796 Prevent spurious "More than one WWN" alerts In previous versions of HP 3PAR OS, 3PAR WWNs which were previously visible on an RCFC port could incorrectly trigger an alert of the form "More than one WWN seen on an RCFC port". The 3.1.2 release of HP 3PAR OS includes modifications to ensure that this alert is raised only as a result of WWNs which are currently visible on the RCFC port. 73837 tunesys no longer requires a Dynamic In previous releases of the HP 3PAR OS, tunesys required a Dynamic Optimization license for system Optimization (DO) license to run. This requirement has been removed rebalancing & systems without a DO license can now be re-balanced. 74520 checkhealth warns about having more vluns than are supported Having more vluns than are supported can result in out of memory issues. The checkhealth command will now flag a warning when the number of luns is beyond the supported limit. 74917 Address deletion of SCSI-3 PGR registrations When more than 16 initiators were registered for SCSI-3 persistent reservations on a single VV, deletion of registered initiators on VV deletion or manual clearing of reservations could introduce inconsistency into the system's Table of Contents. This software defect has been addressed to handle deletion of registered initiators correctly. 75245 Default NL CPGs will use FC drives The default R1 and R6 NL CPGS created by the system will now use for SA space FC drives (if they are available) for SA space. 75839 (also 70918) 3.1.1 MU1 : tunesys fails to distribute chunklets evenly An issue has been reported where the LD tuning phase of tunesys can, in some circumstances, cause an out of space situation. This has been resolved by compacting any residual free space in the related CPG after an LD has been tuned. 76022 77304 77440 System management application restart when starting a region move An issue has been reported where the system management application could restart if thin persistence space reclaim started at the same time as a region move (tune) operation. This issue has been fixed in release 3.1.2. Changes to hw_cage_sled and sw_cage_sled New sw_cage_sled component supports C:M:D format in PD alerts. Existing hw_cage_sled component reverted to support Magazine: x format in HW cage alerts. Fibre Channel driver issue The Emulex driver didn't clean up its device database entry completely when a host hba port was re-assigned to a different FCID by the fabric switch, this led to a stale device entry remaining in the device database. A further look-up in the device database returned a stale 14 HP 3PAR OS 3.1.2 GA Release Notes

  • 1
  • 2
  • 3
  • 4
  • 5
  • 6
  • 7
  • 8
  • 9
  • 10
  • 11
  • 12
  • 13
  • 14
  • 15
  • 16
  • 17
  • 18
  • 19
  • 20
  • 21
  • 22
  • 23
  • 24
  • 25
  • 26
  • 27
  • 28
  • 29
  • 30
  • 31
  • 32
  • 33
  • 34
  • 35
  • 36
  • 37
  • 38
  • 39
  • 40
  • 41
  • 42
  • 43
  • 44
  • 45
  • 46
  • 47
  • 48
  • 49
  • 50
  • 51
  • 52
  • 53
  • 54
  • 55
  • 56
  • 57
  • 58
  • 59
Description
Item
Issue ID
In previous HP 3PAR releases, when a VV tune is cancelled the
displayed VV state remained at ""tuning"" indicating that the tune
VV state returns to "normal" if a tune
is cancelled
70797
must be completed (re-run) or rolled back in order to return the VV
state to normal. This behavior changes in 3.1.2 as the VV state returns
to ""normal immediately on task cancellation.
Avoid potential sysmgr failures by disabling reclaim for a volume if
the total number of vlmaps exceed a certain threshold.
sysmgr panic - too many chunklet
regions specified
71127
When a node is rebooted before the system is fully started, we will
now properly report its status in shownode.
Node state properly indicated after
early node reboot
71684
In rare cases a system that was trying and failing to rebalance LD
node ownership while using Peer Motion to import volumes from
Potential for interaction between LD
node balancing and importvv
operations
72505
another array could get stuck while doing the final VV block to finish
the import. This was a volume block locking issue and has been
resolved to prevent this negative interaction.
A type of meta data inconsistency when discovered caused the node
to panic rather than alert the user via the raising of the Internal
Consistency Event.
Meta Data Inconsistency in bitmaps
can cause node to panic
72983
The servicehost list command will now have greatly reduced memory
consumption and now focuses on the important information.
serivcehost list performance improved
73067
In previous versions of HP 3PAR OS, 3PAR WWNs which were
previously visible on an RCFC port could incorrectly trigger an alert
Prevent spurious "More than one
WWN" alerts
73796
of the form "More than one WWN seen on an RCFC port". The 3.1.2
release of HP 3PAR OS includes modifications to ensure that this alert
is raised only as a result of WWNs which are currently visible on the
RCFC port.
In previous releases of the HP 3PAR OS, tunesys required a Dynamic
Optimization (DO) license to run. This requirement has been removed
& systems without a DO license can now be re-balanced.
tunesys no longer requires a Dynamic
Optimization license for system
rebalancing
73837
Having more vluns than are supported can result in out of memory
issues. The
checkhealth
command will now flag a warning when
the number of luns is beyond the supported limit.
checkhealth warns about having
more vluns than are supported
74520
When more than 16 initiators were registered for SCSI-3 persistent
reservations on a single VV, deletion of registered initiators on VV
Address deletion of SCSI-3 PGR
registrations
74917
deletion or manual clearing of reservations could introduce
inconsistency into the system's Table of Contents. This software defect
has been addressed to handle deletion of registered initiators correctly.
The default R1 and R6 NL CPGS created by the system will now use
FC drives (if they are available) for SA space.
Default NL CPGs will use FC drives
for SA space
75245
An issue has been reported where the LD tuning phase of tunesys
can, in some circumstances, cause an out of space situation.
3.1.1 MU1 : tunesys fails to
distribute chunklets evenly
75839
(also
70918)
This has been resolved by compacting any residual free space in the
related CPG after an LD has been tuned.
An issue has been reported where the system management application
could restart if thin persistence space reclaim started at the same time
System management application
restart when starting a region move
76022
as a region move (tune) operation. This issue has been fixed in release
3.1.2.
New sw_cage_sled component supports C:M:D format in PD alerts.
Existing hw_cage_sled component reverted to support Magazine: x
format in HW cage alerts.
Changes to hw_cage_sled and
sw_cage_sled
77304
The Emulex driver didn't clean up its device database entry completely
when a host hba port was re-assigned to a different FCID by the fabric
Fibre Channel driver issue
77440
switch, this led to a stale device entry remaining in the device
database. A further look-up in the device database returned a stale
14
HP 3PAR OS 3.1.2 GA Release Notes