HP 3PAR OS 3.1.2 MU5 Release Notes

DescriptionItemIssue ID
When resynchronizing a volume over 2TB in size to a rcopy secondary
system, a 32bit counter used to represent the number of contiguous
2.3.1 MU1: sysmgr panicked, system
pending pfail recovery
54956
blocks which has ""delta"" data since the last resync may overflow.
HP 3PAR OS has since been enhanced to use 64bit counters.
Addresses an issue in earlier software releases where Emulex HBA
ports configured in host/fabric mode could be switched to initiator
mode after node replacement.
Configure fabric ports in target mode
after node replacement
55196
In 3.1.1.MU1, there will be IMC support for the scheduler because
this bugfix enables the required functionality in the HP 3PAR OS.
Allows IMC support for scheduler56781
The command setvv -snp_cpg (remove the snapspace CPG) was
causing the System Manager to crash for Virtual Volumes created via
sysmgr panic and degraded single
esi port for all cages when removing
vv from cpg
57495
createaldvv, which had space reclaimed. This command should
now work without error, provided that there is no snapshot on the
Virtual Volume.
To avoid potential problems with stalled remote copy traffic, in
2.3.1.MU5 the creatercopytarget and admitrcopylink
Remote copy traffic may stall if
destination link IP address
corresponds to a local link address
58251
commands will not permit destination link IP addresses that correspond
to local RCIP link addresses or the address of the management
interface.
Media failures on certain chunklets would sometimes prevent relocation
of other chunklets from the raid set as well as returning a preserved
LD back to normal. This has now been fixed.
Fixed chunklet relocation after media
failures.
60029
When a controller node rejoins cluster, HP 3PAR OS attempts to
rebalance the load across system. During the rebalancing process,
HP 3PAR system Panic - tpd:
Assertion point: file: cmp.c, line:
16883
60141
the current HP 3PAR OS does not handle the out of resource condition
which can cause entire system to be offline. The fix is to make sure
there is enough resource before starting the load rebalancing process.
There is a new database created in the “Atlantis" instance, called
"OAuth". Notice that there is a "space" in the name of the database.
Failed snapshots for atlantis cluster -60272
For some reason, definitely a bug with the SQL writer, the writer fails
to come up; which explains why we don't see it in the list of vssadmin
writers. This topic has been discussed, and the user apparently found
the solution with no acknowledgement from Microsoft. http://
social.technet.microsoft.com/Forums/en-US/dpmsqlbackup/thread/
432b4e38-6c7c-4fe0-abe5-769275784fcf/
If we have a heavily loaded system and have snapshots with lots of
data, the meta data tables which are heavily populated, then on
When deleting a very large and
filled snapshot can lock up a CPU for
long time
60494
deleting some of these snapshots the meta data tables are traversed
causing the cpu on occasions to be held for an excessive amount of
time. This can cause the locking out other processes/tasks from running
and can cause the background internode comms processing to be
delayed resulting in the erroneous belief that one of the other nodes
has died. This fix prevents the cpu from being held for an excessive
amount of time while traversing the the meta data tables during
snapshot removal. This fix includes an enhancement (65646) to cover
the case where the meta data tables are being merged during
snapshot removal. This fix with enhancement is applied to 2.3.1.MU5,
3.1.1.MU1, 3.1.2. The fix without the enhancement is in 3.1.1.(This
is also logged as # 52911).
Excessive raw space alerts have been seen when AO is configured
on a system. In 3.1.2 we have provided the ability to set raw space
alert threshold to 10g to suppress these alerts.
Raw space alerts61058
It was possible to try to remove a VV that had been just removed from
an exported VV set before the VV export was fully cleaned up. In
some instances this would cause sysmgr to crash and restart.
Removal of volume from exported VV
set could cause sysmgr crash
61363
Modifications to the OS 9