EMC AX100 Release Notes - Page 15

Fixed problems Problem AX100SC, AX100SCi, AX100, AX100i - Persistent Reservation type not handled properly (according to ANSI spec) can cause Sun host systems to crash. (117505, 117577) AX100SC, AX100SCi, AX100, AX100i - SP panic (0x00800007) during software upgrade (120991) AX100SC, AX100SCi, AX100, AX100i - One or more virtual disks and disk pools disappear after an SP reboots. (119661, 119105, 117351, 117665, 118887, 121446) AX100SC, AX100SCi, AX100, AX100i - Equalize completed on a hot spare but a storage-system reboot or panic during swap-out can lead to a stuck swapped-in hot spare. (114643 AX100SC, AX100SCi, AX100, AX100i - The inmemory copy of the FRU-signature cache can be corrupted resulting in a panic triggered by the corrupted data. (119352) AX100SC, AX100SCi, AX100, AX100i - Internal FLARE GLD configuration file can be deleted because of errors parsing XML file. (119609) AX100SC, AX100SCi, AX100, AX100i - During the Navisphere poll cycle, the timing of an initiator logout can result in ERROR_DEV_NOT_EXIST being returned from an IOCTL that gets logged to the event logs. This is not an error condition. (120086) AX100SC, AX100SCi, AX100, AX100i - Drive removal occurring just as a virtual disk assignment is taken off the disk pool busy queue causes the panic 0x008000e2. (125125) Details Fixed in version: 02.19.100.5.005 Solution: Fixed Persistent Reservation check for non-I/O commands for various reservation types to return the correct status. Fixed in version: 02.19.100.5.005 Solution: There was a very small window of opportunity for an SP to panic during an upgrade if a disk is found dead, is removed, or if a FRU signature error occurs. The code did not wait for the rebuild, which resulted in the virtual disks becoming "broken" and causing the panic. Fixed the code to handle the rebuild properly in compatibility mode during upgrade. Fixed in version: 02.19.100.5.005 Solution: This happened only after a disk pool create failed. One or more virtual disks and disk pools on the same disks could disappear after an SP rebooted. Fixed the code to clean up properly after the failed disk pool creation to avoid the problem. Fixed in version: 02.19.100.5.005 Solution: Swap events are now prevented from occurring while a powerup event is in progress. Also, additional logic has been added to handle a hot spare that is left swapped in after an equalize operation. Fixed in version: 02.19.100.5.005 Solution: Whenever a FRU signature is read, all the partitions are checked for valid virtual disk and disk pool values. Any corrupt partition data is invalidated. Other checks have been added to protect FRU signature data integrity. Fixed in version: 02.19.100.5.005 Solution: A check has been added to ensure that GLD configuration file read operations are successful so future configuration changes do not result in a loss of valid information. Fixed in version: 02.19.100.5.005 Solution: This timing scenario has been accounted for and will return valid status if this window is encountered. Fixed in version: 02.19.100.5.005 Solution: The assign will now fail and be retried. EMC AX100-Series Integrated Management Software Release Notes 15