HP 1606 HP StorageWorks Fabric OS 6.3.1c Release Notes (5697-0509 September 2 - Page 88

Switch panic or inability to execute any command may be experienced after shared memory segments or memory shortage. It can be exposed via: • Creating and removing logical switch multiple times • Performing configdownload with a large configuration file or repeating configdownload • Running repeated perfresourceshow commands with option eeRes or fportRes • Launching port admin and selecting GbE port tab and keeping it open for long time (2 hours) • Cald being busy and not responding back to rpcd for some xml requests causing stale shared memory segments Fixed in Fabric OS 6.3.1b If backbone and edge fabric have different switch PID formats, such as when edge fabric is in IM2 mode while backbone fabric is not, importing another device into the backbone will fail with [FCR-1022] "Local proxy device slot entries exhausted" in raslog. Both routers and switches configured for IR are impacted, and this issue is more likely to be seen if there are frequent device offline/online transitions in edge fabrics. Fixed in Fabric OS 6.3.1b Unable to download Fabric Watch settings to the switch when mismatched names (LinkReset/LReset) are used in the configuration file. DCFM unable to replicate Fabric Watch setting to switch and fails with raslog error. Occurs only with linkreset setting in configuration files. Fixed in Fabric OS 6.3.1b Workaround prior to upgrade: Once switch is disabled, the configuration downloaded and the switch re-enabled, the configuration can be modified and downloaded without any problems. Can also remove or comment out the link reset settings in the .cfg file. Switch resets under very rare condition of lost heartbeat while active CP is running Fabric OS 6.3.x and standby CP is running 6.2.x during Fixed in Fabric OS 6.3.1b firmware download. Unexpected reboot and failover when attempting to push zone DB from DCFM in IM2/IM3 fabrics. Fixed in Fabric OS 6.3.1b Workaround prior to upgrade: Use CLI or EFCM for non-TI/redirect zones) in IM2/IM3. In a routed fabric environment (routers and/or switches running IR mode), WWN for a device shows up in multiple edge fabrics. Proxy devices for the same host/target could be created in unintended fabric which could cause traffic to be sent to the wrong fabric and the host not seeing the target. Occurs when there is frequent F port bounce (12 sec) in FCR fabric. Issue exists only in Fabric OS 6.3.x. Fixed in Fabric OS 6.3.1b 4Gb director does not respond to Link Reset primitives correctly. Due to timing condition after reboot or switchdisable and switchenable, or when device detects credit shortage, devices sent LR to the director but it does not respond and the link goes down. Fixed in Fabric OS 6.3.1b Access Gateway NPIV port gets disabled as "initiator/target connected on N_Port" after upstream encryption switch was rebooted, or encryption blade executes a Power On Self Test. When bpportloopbacktest is run as part of a POST on the encryption switch or encryption blade, loop back mode caused the FLOGI frames from the AG ports on the AG switch to loopback and further disable those ports. Fixed in Fabric OS 6.3.1b Workaround prior to upgrade: Disable POST on encryption switch or chassis with encryption switch, or reenable the AG ports manually. 88