HP A7533A - Brocade 4Gb SAN Switch Base Release Note page 70

sysshutdown on 8Gb SAN Switch for HP BladeSystem c-Class may
cause panic and reboots.
HP StorageWorks Multiprotocol Router Blade heartbeat stops, causing
a reboot while traffic is running over FCIP links. This may cause CP
failover, disrupting traffic only on the router blade.
After changing the domain offset followed by hafailover on CP1
in interopmode 2 on DC SAN Director, DC04 Director or 4/256 Dir-
ector, Unknown LSR type messages generated by CP0 scroll con-
tinuously on console log.
Under heavy traffic on 10-Gb tunnel on HP StorageWorks DC SAN
Director Multiprotocol Extension Blade running in compression mode,
back-to-back tunnel parameter modification can cause FCIP traffic to
halt due to timing issue with internal buffers. This can also happen on
1-GbE ports, but over a longer I/O run than on 10-GbE ports.
I/O times out and stops when running across same VE link as other
I/O when failover is done on a trunked EX_Port with HP Storage-
Works 1606 Extension Switch or Multiprotocol Extension Blade.
Zoning fails with Transaction Commit failed. Reason code
2 (26) - Aca was rejected: Remote Switch Busy, Retry
in a few seconds error when E_Port to neighboring Mi10k is lost
during a zone push operation when in interop mode.
I/O traffic with 1606 Extension SAN Switch or Multiprotocol Extension
Blade unable to run across edge-to-edge topology when compression
and FastWrite are enabled in certain Virtual Fabric configurations.
SFP power-supply values are not displayed correctly after doing a
nondisruptive firmware upgrade to Fabric OS 6.2.1.
FCIP tunnel may bounce up and down in an impairment network con-
figuration with compression enabled and highly compressible data
I/O with either the 1606 Extension SAN Switch or Multiprotocol Ex-
tension Blade.
With either the 1606 Extension SAN Switch or Multiprotocol Extension
Blade, VE tunnel did not come up due to configuration mismatch for
tape pipelining. One side shows Write only and the other side
shows Write/Read.
High CPU load from DCFM performing supportsave causes port to
go offline with laser FLT/link reset.
Name server database counts do not match for Fabric OS (6.2.1) and
M-EOS (9.9.0) switches in the same fabric, so host cannot see the
device. This occurs when an HBA capable of a FDMI Register HBA
(RHBA) command and M-EOS switch are in the same fabric.
Switch panic triggered by size-64 kernel message buffer being used
up in virtual fabric with logical ISL configured.
70
Fixed in Fabric OS 6.3.0a
Workaround prior to code change:
Use the CLI shutdown command.
Fixed in Fabric OS 6.3.0a
Fixed in Fabric OS 6.3.0a
Workaround prior to code change:
Perform hafailover to other CP.
Fixed in Fabric OS 6.3.0a
Fixed in Fabric OS 6.3.0a
Fixed in Fabric OS 6.3.0a
Fixed in Fabric OS 6.3.0a
Fixed in Fabric OS 6.3.0a
Workaround prior to code change:
Execute the command again to obtain
the correct values.
Fixed in Fabric OS 6.3.0a
Fixed in Fabric OS 6.3.0a
Fixed in Fabric OS 6.3.0a
Fixed in Fabric OS 6.3.0a
Fixed in Fabric OS 6.3.0a
Workaround prior to code change:
Configure only DISLs between
switches.