EMC AX100 Release Notes - Page 13

Fixed problems Hardware problems fixed in version 02.19.100.5.005 Problem AX100SC, AX100SCi, AX100, AX100i Replacing a drive that is being equalized while the disk pool is shut down may lead to incorrect data on the rebuilt disk. This can lead to a CM_ILLEGAL_FRU_STATE (0x00800064) panic. (121839) AX100SC, AX100SCi, AX100, AX100i - In rare circumstances, incorrect data is written to an expanded virtual disk if media errors are recorded on one of the drives in the disk pool. (123356) AX100SC, AX100SCi, AX100, AX100i - In cases where there are multiple uncorrectable errors, data verification may result in repeated panics or invalidation of otherwise recoverable data. (123526) AX100SC, AX100SCi, AX100, AX100i - A storage system experienced a dual-SP panic while rebuilding a RAID 1/0 mirror to a hot spare. This may cause a data unavailable situation. (123867) AX100SC, AX100SCi, AX100, AX100i - If virtual disks within a disk pool are trespassed while an equalize is in process, the replacement disk may contain incorrectly rebuilt data. (123878) AX100SC, AX100SCi, AX100, AX100i - Under certain circumstances, multiple drive failures within a disk pool can cause unowned virtual disks during a bind operation while a drive in the disk pool is in the process of being rebuilt to a hot spare. This may cause a data unavailable situation. (124079) AX100SC, AX100SCi, AX100, AX100i - Disk drives that were included in an attempt to create a disk pool are unavailable for use in additional attempts if a disk fails during the create operation. This situation could prevent the user from creating a disk pool until after the SP is rebooted. (122405) Details Fixed in version: 02.19.100.5.005 Solution: Reset equalize checkpoint when a new disk is inserted in the slot being equalized, and rebuild the FRU database when swapping out a disk. Fixed in version: 02.19.100.5.005 Solution: A problem was uncovered in and fixed in the error handling code. Fixed in version: 02.19.100.5.005 Solution: This requires multiple errors to occur, leading to inconsistent metadata. Changes have been made to have the metadata updated such that reconstruction of data can occur. Fixed in version: 02.19.100.5.005 Solution: A change was made to hold off the equalize until all virtual disks have finished rebuilding. Fixed in version: 02.19.100.5.005 Solution: Changes were made to the RAID driver to use the most accurate equalize checkpoint information. Fixed in version: 02.19.100.5.005 Solution: The bind operation was modified to include a check to make sure that it does not mark a virtual disk on a device as needing a rebuild if another device in the disk pool is already marked. This change will prevent the double fault situation. Fixed in version: 02.19.100.5.005 Solution: The peer SP prevents the use of the disks for subsequent disk pool creation attempts because it thinks that the disk pool exists. The software was changed so that the peer SP would not recognize the disk pool as created if the operation failed. Prior to this release, the workaround for this problem required rebooting the SP (peer SP) that thinks the disk pool exists. EMC AX100-Series Integrated Management Software Release Notes 13