HP 3PAR StoreServ 7400 2-node HP 3PAR OS 3.1.2 Release Notes - Page 13

Promote XR3B drive FW for Moose, Improved UNMAP performance

Page 13 highlights

Issue ID Item Description to facilitate expedited synchronization post migration. This functionality was integrated into the createsv command, using the -rcopy switch. 67556 Help for createsv CLI command does createsv now alows read-write snapshot creation from read-write not reflect changes to functionality base/read-write snapshots 67913 Firmware issue could cause port resets due to mailbox command timeouts Upgraded 1G iSCSI firmware to version 3.0.1.67 in order to address a firmware issue which could cause port resets due to mailbox command timeouts under heavy I/O load. 67983 Secondary iSNS address removed Secondary iSNS address has been removed from CLI since it was from CLI never supported. 68750 tunesys - added support for AO volumes The system tuning command tunsys now supports Adavtive Optimisation (AO) configured volumes split between CPGs. 68952 New alert if a 1G iSCSI port comes up at a lower speed If a 1G iSCSI port on the array comes up at a negotiated port rate that is lower than the expected 1G rate, an alert is generated. The showport -iscsi command now displays the actual rate; previous releases used to display 1Gbps regardless of the actual rate the link came up at. 69011 Fixed manual 8Gps rate setting for RCFC & PEER ports This release fixes the controlport rate command to allow for a manual rate setting of 8Gps on rcfc and peer ports that are capable of that data rate. Previously only the rate setting of 'auto', which is the default, would achieve a 8Gps rate for those ports. 69130 Cage Inaccessible errors after master Misleading alerts indicating that a cage is inaccessible, are no longer node/sysmgr panic emitted at the time of System Manager startup. 69193 Fibre Channel driver issue Fixed a driver bug that could cause a link list break in a very rare condition. 69439 Deprecate updatevv's -anyid option The updatevv -anyid option is deprecated. In 3.1.2 we changed the volume and snapshot removal handling in System Manager. When a volume or snapshot is removed, it is not cleaned up immediately. It remains hidden until a background thread recycles the resource. That means upon removal, the id of the volume is still being occupied, so updatevv may fail. In 3.1.2, we are using a different volume id but keep the same WWN when the snapshot is updated. Now updatevv calls a modified creategroupsv command which will always use the same WWN to create the new snapshot. 69556 Promote XR3B drive FW for Moose drives The XR3B firmware will lower a temperature threshold. The patch will modify the firmware database to cause the required firmware on Moose drives to be XR3B. SN: P90 will modify the scsidb to point the current firmware for Moose to be XR3B. As this is a scripted modification the scsidb it will need to be applied after every upgrade, before admithw is run. 69606 Pinned data after growing a VV part In cases where NL drives are overloaded in the extreme, upper layers of a RC group can see this as loss of access. If multiple NL drives are in this state, this loss of access can result in volumes going into preserved state. Along with other fixes to prevent this type of overloading (as in 77976) we also now allow longer NL access times and ensure multiple retries to ensure what looks like loss of access is actually so. 70121 Network issue when node is added/removed If a network message is sent to a node that then is removed from the system, the networking infrastructure hangs. Any future network issue cannot be handled, ultimately resulting in connectivity loss to the cluster until the network interface is restarted. A fix was made so that the network interface does not hang in this situation but eventually times out instead, allowing handling of other networking changes. 70425 Improved UNMAP performance Unmap SCSI commands issued to already-unmapped logical block address ranges of thinly-provisioned volumes are handled more efficiently in HP 3PAR OS 3.1.2. This can lead to improved performance during formatting operations for some host types. Modifications to the OS 13

  • 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
to facilitate expedited synchronization post migration. This functionality
was integrated into the
createsv
command, using the -rcopy switch.
createsv now alows read-write snapshot creation from read-write
base/read-write snapshots
Help for createsv CLI command does
not reflect changes to functionality
67556
Upgraded 1G iSCSI firmware to version 3.0.1.67 in order to address
a firmware issue which could cause port resets due to mailbox
command timeouts under heavy I/O load.
Firmware issue could cause port
resets due to mailbox command
timeouts
67913
Secondary iSNS address has been removed from CLI since it was
never supported.
Secondary iSNS address removed
from CLI
67983
The system tuning command
tunsys
now supports Adavtive
Optimisation (AO) configured volumes split between CPGs.
tunesys - added support for AO
volumes
68750
If a 1G iSCSI port on the array comes up at a negotiated port rate
that is lower than the expected 1G rate, an alert is generated. The
New alert if a 1G iSCSI port comes
up at a lower speed
68952
showport -iscsi
command now displays the actual rate; previous
releases used to display 1Gbps regardless of the actual rate the link
came up at.
This release fixes the
controlport rate
command to allow for a
manual rate setting of 8Gps on rcfc and peer ports that are capable
Fixed manual 8Gps rate setting for
RCFC & PEER ports
69011
of that data rate. Previously only the rate setting of 'auto', which is
the default, would achieve a 8Gps rate for those ports.
Misleading alerts indicating that a cage is inaccessible, are no longer
emitted at the time of System Manager startup.
Cage Inaccessible errors after master
node/sysmgr panic
69130
Fixed a driver bug that could cause a link list break in a very rare
condition.
Fibre Channel driver issue
69193
The
updatevv -anyid
option is deprecated. In 3.1.2 we changed
the volume and snapshot removal handling in System Manager. When
Deprecate updatevv's
-anyid
option
69439
a volume or snapshot is removed, it is not cleaned up immediately.
It remains hidden until a background thread recycles the resource.
That means upon removal, the id of the volume is still being occupied,
so
updatevv
may fail. In 3.1.2, we are using a different volume id
but keep the same WWN when the snapshot is updated. Now
updatevv
calls a modified
creategroupsv
command which will
always use the same WWN to create the new snapshot.
The XR3B firmware will lower a temperature threshold. The patch will
modify the firmware database to cause the required firmware on
Promote XR3B drive FW for Moose
drives
69556
Moose drives to be XR3B. SN: P90 will modify the scsidb to point the
current firmware for Moose to be XR3B. As this is a scripted
modification the scsidb it will need to be applied after every upgrade,
before admithw is run.
In cases where NL drives are overloaded in the extreme, upper layers
can see this as loss of access. If multiple NL drives are in this state,
Pinned data after growing a VV part
of a RC group
69606
this loss of access can result in volumes going into preserved state.
Along with other fixes to prevent this type of overloading (as in 77976)
we also now allow longer NL access times and ensure multiple retries
to ensure what looks like loss of access is actually so.
If a network message is sent to a node that then is removed from the
system, the networking infrastructure hangs. Any future network issue
Network issue when node is
added/removed
70121
cannot be handled, ultimately resulting in connectivity loss to the
cluster until the network interface is restarted. A fix was made so that
the network interface does not hang in this situation but eventually
times out instead, allowing handling of other networking changes.
Unmap SCSI commands issued to already-unmapped logical block
address ranges of thinly-provisioned volumes are handled more
Improved UNMAP performance
70425
efficiently in HP 3PAR OS 3.1.2. This can lead to improved
performance during formatting operations for some host types.
Modifications to the OS
13