HP 3PAR InForm OS 3.1.1 MU1 Release Notes

DescriptionItemBug ID
A problem in the code used to detect the state of node sensor could cause a
node to enter a Degraded state incorrectly. This has been fixed in InForm OS
3.1.1 MU1
Node sensor fails and
recovers intermittently
66592
Host IO during heavy internal defrag load could result in a system hang. This
has been fixed in InForm OS 3.1.1 MU1.
Internal resource
mismanagement can
66806
cause system hang
under heavy defrag
load
A problem was found in log handling related to space deallocation when a
node rejoins 2-node cluster. As a result, volume meta data could become
Mishandling of logs
when a node rejoins
66807
inconsistent and uncontrolled node-down can be triggered. This has been fixed
in InForm OS release 3.1.1 MU1.
cluster in a 2-node
system
The user and snap space defragmentation can consume a lot of system resources
in short bursts and can cascade into leaving the Storage System unresponsive
to host I/O. To address this issue, space defragmentation has been throttled.
Throttle user/snap
space defragmentation
66808
To improve security, some cipher and MAC (Message Authentication Code)
algorithms have been disabled. We now allow only aes192-ctr, aes256-ctr,
Allowed encryption
methods for SSH
reduced
66810
aes128-ctr, aes192-cbc, aes256-cbc, aes128-cbc, and 3des-cbc for the
encryption algorithm. We allow only hmac-sha1 and hmac-sha1-96 for the MAC.
An issues in the HBA driver is causing the FC044X card's Qlen to be set to the
LP11002 card's Qlen.
HBA driver settings67095
Reading or accessing the audit log is, itself, an auditable event. Hence, invoking
showeventlog either implicitly or explicitly now adds a getevenlog entry in
the event log.
Logging
showeventlog
requests
67444
Due to an overflow, sysmgr would panic when checking the user allocation limit
for checkhealth if the volume was 2TB or greater in size. To address this, the
Checkhealth allocation
limit may cause a
sysmgr panic
67803
type used for the computation has increased in size to prevent other potential
overflows.
Due to a defect in InForm OS 3.1.1, if PDs are not present on all nodes, flusher
on the nodes that do not have PD will not work properly. This can cause dirty
cache pages to accumulate which impacts performance.
Flusher does not work
properly on nodes with
no PD connections
68039
After an uncontrolled shutdown, logical disks belonging to nodes which do not
come up could remain unstarted. Fixed the logical disk startup code to handle
this situation properly.
Fixed a condition that
was causing LDs to
remain unstarted after
68205
an uncontrolled system
shutdown
Fixed the chunklet relocation to fail if there is double disk failure with RAID-5 or
triple disk failure with RAID Multi-Parity (MP).
The movech command
should fail on RAID-5
69470
double-disk failure or
RAID-6 triple disk
failure
34 InForm OS 3.1.1 MU1 Release Notes