HP StorageWorks 2/64 HP StorageWorks Fabric OS 5.0.5d Release Notes (AA-RW1WF- - Page 9

Fabric OS 5.0.5b fixes Table 3 lists improvements included in Fabric OS 5.0.5b. Table 3 Fabric OS 5.0.5b closed defects Closed Defects Inserting unsupported FC4-48 blade into a pre-Fabric OS 5.2 chassis causes FAULTY(1) error that will not clear with blade removal. HAsync is dropped, and an active CP reboot is required to recover. If not upgrading to this Fabric OS 5.0.5b release the workaround is to schedule a convenient time to reboot the active CP. HBA occasionally may not log into a 4Gb switch during auto topology discovery. If not upgrading to this Fabric OS 5.0.5b release the workaround is to use portCfgGPort to lock the switch in G port mode. To reestablish the link on the affected port, issue portDisable and portEnable. Switch panics with Failure Detection (FDET) when a Linux host is connected, affecting the SAN Switch 2/32, SAN Director 2/128, SAN Switch 2/8V and SAN Switch 2/16V platforms. A phantom device on an L-port disappears after HAfailover/hareboot, and a private device is no longer in the name server database, affecting the SAN Switch 2/32, Core Switch 2/64, SAN Director 2/128, SAN Switch 2/8V and SAN Switch 2/16V platforms. If not upgrading to this Fabric OS 5.0.5b release the workaround is to issue the portdisable and portenable commands. Specific HBA behavior causes a memory leak and the switch runs out of memory and reboots. Switch panics with time service daemon (tsd) core dump. Zoning change fails with message: unabortable internal zoning transaction. Zoning change fails with message: unabortable internal zoning transaction. The Web Tools GUI hangs upon invoking fabricwide polling when there is no IP connection between the Web Tools proxy switch and the other switches in the fabric. Without displaying an error message, a hardware ASIC failure causes performance degradation on a 4/256 SAN Director, or a switch fault on other 4Gb switches. If not upgrading to this Fabric OS 5.0.5b release the workaround is to replace the faulted blade. During stress testing, after fastboot on a fully-loaded 4/256 SAN Director a few E_ports and F_ports may stay in "Mod_Val" mode. If not upgrading to this Fabric OS 5.0.5b release the workaround is to disable and enable the port that is stuck in the Mod-Val state. During 3rd party SFP qualification testing with certain types of SFP, a high number of cyclic redundancy check (CRC) and low level interrupt (LLI) errors were reported by porterrshow. On the Core switch 2/64, logging into both logical IP addresses puts the user on the same logical switch. If not upgrading to this Fabric OS 5.0.5b release the workaround is to use ssh (secure shell) or after logging in as root, change the logical switch instance: Enter the setswitch command where is either 0 or 1. Change 0 to 1 if needed. The PS1 line needs to be entered only once per login. Port blade on a SAN Director 2/128 faulted due to Failure Detection when there is a 1Gb F_port on an even-numbered user port. If not upgrading to this Fabric OS 5.0.5b release the workaround is to connect 1Gb F_ports to odd-numbered user ports only. HP StorageWorks Fabric OS 5.0.5d release notes 9