OPERATIONAL DEFECT DATABASE
...
...
Scenario During read and write IO operation using the Toshiba model KPM5XVUG1T92 mixed use SSD devices and model KPM5XRUG3T84 read intensive SSD devices multiple timeouts occur on many different SDSs simultaneously. Causing double failure and a DATA_FAILED state on many copies of data. Note that this affects all Toshiba KPM5X series models and not just the ones noted here, these two models are commonly sold and used with VxFlex systems and were the ones found to be problematic in multiple instances. Symptoms Data unavailability happens , and many devices show an error state like the below figure. From the Linux message logs we can see the following CBD entries: kernel: scsi_io_completion: 4 callbacks suppressed kernel: sd 0:2:2:0: [sdc] tag#0 FAILED Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK kernel: sd 0:2:2:0: [sdc] tag#0 CDB: Read(10) 28 00 00 00 00 00 00 00 08 00 kernel: blk_update_request: 4 callbacks suppressed kernel: blk_update_request: I/O error, dev sdc, sector 0 kernel: sd 0:2:2:0: [sdc] tag#0 FAILED Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK kernel: sd 0:2:2:0: [sdc] tag#0 CDB: Read(10) 28 00 00 00 00 00 00 00 08 00 kernel: blk_update_request: I/O error, dev sdc, sector 0 kernel: Buffer I/O error on dev sdc, logical block 0, async page read From any PERC RAID controller term logs you see multiple sense code 5/24/0 and CMD timeouts: 09/03/19 15:37:35: C0:EVT#04033-09/03/19 15:37:35: 113=Unexpected sense: PD 02(e0x40/s2) Path 12000000, CDB: 12 01 c0 00 20 00, Sense: 5/24/00 09/03/19 15:37:35: C0:RS PD 2: 70 00 05 00 00 00 00 28 00 00 00 00 24 00 00 00 00 00 00 12 40 18 01 40 00 00 00 00 00 00 79 00 00 1d 33 19 18 02 02 41 00 00 00 00 00 00 00 00 09/03/19 15:37:35: C0:EVT#04034-09/03/19 15:37:35: 113=Unexpected sense: PD 02(e0x40/s2) Path 12000000, CDB: 12 01 c0 00 20 00, Sense: 5/24/00 09/03/19 15:37:35: C0:RS PD 2: 70 00 05 00 00 00 00 28 00 00 00 00 24 00 00 00 00 00 00 12 40 18 01 40 00 00 00 00 00 00 79 00 00 1d 33 19 18 02 02 41 00 00 00 00 00 00 00 00 09/03/19 15:37:35: C0:process_dcdb_callback: No valid issuer Sense address for the sense Data sdAddr:1dba0ac sdLen:0 cmdId:f4 09/03/19 15:37:35: C0:CMD_PCI: cmd=03, cmdId=f4, nsg=1, pd=02, timeout=1e, cdb= 12 01 c0 00 20 00 The RAID controller can also be reset during these issues as well due to the PERC or HBA attempting to recover the failed devices. Impact Data unavailability while multiple devices are in error state on multiple SDSs in the same storage pool.
Currently there is a bug in the Toshiba B018 FW version for the PM5 drives that causes them to have command timeouts during IO.
To resolve this upgrade all Toshiba PM5 devices to B01A or B01C firmware version. Then clear all device errors. PM5 B10C FW Download PM5 B10A FW Download Impacted Versions VxFlex All Versions can be affected by this. Fixed In Version Toshiba Drive Firmware B01A/B01C
Click on a version to see all relevant bugs
Dell Integration
Learn more about where this data comes from
Bug Scrub Advisor
Streamline upgrades with automated vendor bug scrubs
BugZero Enterprise
Wish you caught this bug sooner? Get proactive today.
