HP StorageWorks Fabric OS 5.3.1a release notes (AA-RWEYJ-TE, May 2008)

Closed defect summary
Solution
With large z ones
et (over 1200 zones) A PI library crashes as the zone
library does memcpy that does not have terminating zero in the buffer.
Fixed in Fabric OS 5.3.0
Change made in 5.2.0 to limit the number of F DISCs per un it of time,
caused timeouts of NPIV logins in some BladeServer configurations
with VC-FC.
Fixed in Fabric OS 5.3 . 0.
When FOS version prior to 5.2.0 is upgraded to a later version and
the fabri c.ops.mode.pidFormat value was previously set to z ero, the
switch will continually reboot.
Fixed in Fabric OS 5.3.0
In extremely rare c onditions when a par ticular timing window is hit
at the same time an internal end to en d monitor wraps its counter,
the standby CP can get stuck in a reboot loop. The condition is
that due to
a race condition, a partial HA update arrives before a
full updat
e from the active CP, resulting in standby CP g oing into
continuou
srestartcycle.
Fixed in Fabric OS 5.3.0
WhenMSrequestissuedwithWWNthatdidn’texistinthefabric,
the MS Daemon panicked.
Fixed in Fabric OS 5.3.0
The2/128wasforwardingPLOGIframesreceivedwithDIDas
broadcast format which was c ausing issues with other devices in
the fabric.
Fixed in Fabric OS 5.3.0. No longer
forwardsthesetypeofframes.
For platf
orms running 5.2.0 and above with domain, port zoning,
thereisasmallwindowduringwhichifadevicesendsitPLOGI
very fas
t, and before the nameserver information has been fully
propag
ated, the PLOGI may be dropped.
Fixed in Fabric OS 5.3.0
CLI systemverification command incorrectly reports false
positive failure on FC4-16IP Blade. This diagnostic command is not
typically run by a customer.
Fixed in Fabric OS 5.3.0
“Stats” menu option d oes not work when executing minicycle
command.
Fixed in Fabric OS 5.3.0
Process login drop due to devices in different edge fabrics with same
node W
WN. Applies to 4/256 with FCR-18i Blade, a nd 400 MP
Rout
er. Routers may not forward the Nameserver request to proper
doma
in, resulting in request timeout & retries every 2 seconds.
Fixe
d in Fabric OS 5.3.0
In a FICON environment a Switch panic (in the Management Server
daemon) could be seen if the switch receives one of three FI CON
specific frames (FICON GCAP, RLIR, RNID) and the fabric goes
"unstable" ( fabric being rebuilt such as when one of the switches in
the fabric is powered off) before a response is sent.
Fixed in Fabric OS 5.3.0
Traffic can’t be routed into backbone fabric from EX_Port on 400 MP
RouterwithdomainID7or12andfromFR4-18iBladein4/256
with various domain ID’s.
Fixed in Fabric OS 5.3.0
Rapidly issuing repeated supportsave commands can cause
s
witch panic due to out of memory condition.
Fixed in Fabric OS 5.3.0
Wheninteropmodesetto1,“DisableNodeNameZoneChecking”
parameteralsochangesfrom0to1,butreturnsto0afterswitch
reboot .
Fixed in Fabric OS 5.3.0
Due to race condition in RNID command from FICON channels,
FICON CUP daemon crashes and switch panics after switch comes
up then it starts processing the RNID command. Problem affects
FICON environment for Fabric OS 5.1 and later.
Fixed in Fabric OS 5.3.0
Web Tools error message provided when fmsmode setting fails is
not clear.
Fixed in Fabric OS 5.3.0
HP StorageWorks Fabric OS 5.3.1a release notes
19