HP 4400 Release Note page 34

Table 8 Fabric OS 6.2.1 closed defects
Closed defect summary
Sshd allows admin login to execute root-level Linux commands, but admin
CLI commands fail with a command not found error.
On rare occasions, an extra RSCN is delivered when HA failover is
performed before the name server is in sync.
Firmware is not enforcing a mandatory switch disable before configuring
the backbone fabric ID.
Common Area Loader Daemon (CALD) shared memory is not released,
leading to high CPU load and inability to run many of the CLI commands;
causes no space left on device error.
In mixed Fabric OS and M-EOSc fabric in interopmode 2, with more than
500 devices, nameserver database counts do not match. This issue does
not exist with M-EOSn based switches in a mixed fabric.
Toggling the N_Port could cause the F_Port to get disabled, leading to the
crypto LUN getting stuck in a Not ready (LUN rediscovery) state with HP
StorageWorks Encryption SAN Switch in Access Gateway mode.
A switch having a zone object with an extra ";" or -p flag for a zone create
does not perform an error check for alphanumeric characters, which can
trigger loss of zone configuration after a firmware upgrade, or can cause
a switch panic after the cfgenable command.
In Access Gateway mode, the FDMI database does not show proper
information after disabling the port connected to an HBA. Brocade_Software
Identity instance for the remaining node is missing in Brocade SMIA.
During a stress test with switchdisable/switchenable of all switches
in a fabric, we observed both verify failure and a switch panic in the RTE
module.
A small timing window exists due to a corner case where, if a device sends
PLOGI before the route is set up, it could cause PLOGI to drop in single
switch fabric. This is caused by race condition where PLOGI is transmitted
prior to the name server query, when the internal route has not yet been
established for the WWN zoning pair.
Authentication fails after 15 minutes if external authentication, such as LDAP
or Radius server, is configured. The user is not able to login using Web
Tools or web page. Invalid password is reported. This issue does not occur
when the CLI is used.
I/O cannot resume when jamming the status frame with IDLE with QoS
enabled. Link reset does not load the buffers correctly and I/O can be
disrupted
A Common Area Loader Daemon (CALD) memory leak occurs while running
management applications using the SMI interface. This causes a switch
panic/restart and fabric disruption.
34
Solution
Fixed in Fabric OS 6.2.1
Workaround prior to update: A
workaround to facilitate execution of
admin level CLI commands using ssh, is
to use syntax similar to ssh -l admin
192.168.1.12 rbash --login -c
switchshow. Or, if a command is
needed that works on both older (correctly
operating) Fabric OS and the
newer/affected Fabric OS: ssh -l
admin 192.168.1.12 switchshow
|| rbash --login -c
switchshow.
Fixed in Fabric OS 6.2.1
Fixed in Fabric OS 6.2.1
Fixed in Fabric OS 6.2.1
Workaround prior to update: Run
hafailover prior to the upgrade.
Fixed in Fabric OS 6.2.1
Fixed in Fabric OS 6.2.1
Fixed in Fabric OS 6.2.1
Fixed in Fabric OS 6.2.1
Fixed in Fabric OS 6.2.1
Fixed in Fabric OS 6.2.1
Fixed in Fabric OS 6.2.1
Fixed in Fabric OS 6.2.1
Fixed in Fabric OS 6.2.1