HP StorageWorks 1606 HP StorageWorks Fabric OS 6.3.0c Release Notes (5697-0361 - Page 74

Closed defect summary Solution When running a third-party host scan, the http daemon (weblinker.fcg) terminats and restarts several times, causing a temporary management service interruption. Fixed in Fabric OS 6.3.0b HP StorageWorks DC SAN Director panics due to a termination of msd during fabric reconfiguration while executing a switchdisable-agshow-switchenable loop in a fabric with an Access Gateway device with more than 56 switches. Fixed in Fabric OS 6.3.0b The SNMP manager gets incorrect MIB information from a switch running Fabric OS 6.1.x due to a change in SwFwEPortUtil and SwFwEPortPktl between SW_v5_7 to SW_v5_8 MIB. Fixed in Fabric OS 6.3.0b In a 2408 FCoE Switch and a 10/24 FCoE Blade, when CNA sends FIP ver=0, the switch should not validate the MAC descriptor field for backwards compatibility, but it does, leading to the switch rejecting the login. Fixed in Fabric OS 6.3.0b Supportsave does not clear the FFDC files, leading to continued Fixed in Fabric OS 6.3.0b FFDC file warnings being issued until a second supportsave is taken. Fabric OS 6.3.0c fixes Table 11 lists defects closed in the Fabric OS 6.3.0c firmware release. Table 11 Fabric OS 6.3.0c closed defects Closed defect summary Solution HP StorageWorks 4/256 SAN Director with multiprotocol router blade does not see a new device across LSAN after upgrading from Fabric OS 5.x non-disruptively to Fabric OS 6.2.x. Bounce of host port causes paths to be lost between the host and target. This problem is relevant only to this particular router blade and only after upgrading to Fabric OS 6.2.x. Fixed in Fabric OS 6.3.0c Workaround prior to upgrade: Do a slotpoweroff/slotpoweron of the router blade or reboot. When integrated routing is running, FCR can fail to route traffic on 8 Gb EX_Ports of HP StorageWorks DC SAN Director after upgrading firmware from Fabric OS 6.1 to 6.2.x. If there is an additional hafailover after upgrading firmware, a new device is added and bouncing device ports can trigger a routing problem, causing the host to not see the device. Fixed in Fabric OS 6.3.0c Workaround prior to upgrade: If problem has already occurred, bounce all EX_Ports in the same chip together. After recovery, fabric is no longer susceptible to the problem. Switch panic occurs when management application performs repeat FCP LUN emulation queries. If the device gives an inappropriate response, or no response, to an FCP LUN request, FCP Daemon retries the query and the thread becomes stuck. Any new FCP LUN request that is issued from the management application generates a new thread. Eventually FCP Daemon generates up to the maximum number of allowed threads, leading to the switch panic. Fixed in Fabric OS 6.3.0c 74