HP 4400 HP StorageWorks Fabric OS 6.2.2e Release Notes (5697-0809, February 20 - Page 35

Table 8 Fabric OS 6.2.1 closed defects (continued) Closed defect summary Solution The following SNMP-related issues exist in code: Fixed in Fabric OS 6.2.1 • The cpLastEvent value is shown incorrectly in the browser. • The cpStatusChange trap is not sent to the trap recipients. • There is an incompatibility with client using older MIBs (pre-Fabric OS 6.2.0 MIB). • The swEndDeviceInvalidWord mid command returns bad words. • SNMP displays incorrect index value for atTable, ipTable, and ipNetTodefia Table of MIB-2. • MIB fails to compile on third-party SAN Management software (swPmgrEventType). After upgrade to Fabric OS 6.2.0, Fabric Watch reports HP StorageWorks Fixed in Fabric OS 6.2.1 4/8, 4/16 SAN switches, and embedded 4-Gb switches as unhealthy and issues a warning message that compact flash is 90% full. A bad SFP triggered an i2C access causing the following problems on the Fixed in Fabric OS 6.2.1 port/slot display and CPU load on HP StorageWorks 4/256 SAN director: • Executing a slotshow command causes a no system control access to slot error. • The sfpshow command gives Mod_val, or an i2C access message. Port statemismatch occurs on internal ports of 4-Gb embedded Fixed in Fabric OS 6.2.1 switches between switchshow and portcfgshow. The portcfgshow command shows a persistent disable state while the port is enabled. After removing EE and F_Port TT filters on HP StorageWorks DC SAN director in non-VF (Virtual Fabrics) mode, and enabling Fabric mode Top Talkers, the flow values are inaccurate. Fixed in Fabric OS 6.2.1 If TI zone operations are performed simultaneously on different switches, Fixed in Fabric OS 6.2.1 the active CP may failover due to termination of the name server module A switch restart on HP StorageWorks DC SAN or DC04 director (power Fixed in Fabric OS 6.2.1 cycle or reboot of both CPs) causes random ICL E_Ports to be segmented. Due to a rare corner case, a port reset at the same time as an end-to-end Fixed in Fabric OS 6.2.1 PLOGI or during a PLOGI storm can lead to a kernel panic with the message Oops: kernel access of bad area, sig: 11 [#1] and sometimes cold recovery. HA failover on HP StorageWorks DC SAN director times out, causing a switch reboot. Fixed in Fabric OS 6.2.1 Workaround prior to update: Manually block/unblock an E_Port to rebuild the fabric. Daemons may panic during FID (logical switch) deletion and creation if there is no significant time delay between the deletion and creation. The switch becomes unresponsive and the daemon panics. Fixed in Fabric OS 6.2.1 In a FICON environment, a switch may panic when blocking/unblocking Fixed in Fabric OS 6.2.1 an E_Port due to incorrect RNID data count. Concurrent reboot of HP StorageWorks DC SAN directors at both ends of Fixed in Fabric OS 6.2.1 DWDM link causes the link to fail to come up, because the link is unable to sync before the port is enabled at the correct speed. Devices attached to a switch in Access Gateway mode are reported by DCFM/weblinker as physical rather than virtual devices. Fixed in Fabric OS 6.2.1 A switch reboot/hafailover with kernel assert in the FCPH module occurs Fixed in Fabric OS 6.2.1 when the received frame size is too big. Fabric OS 6.2.1 fixes 35