FC LTO4 drive: Info fld=0xfc00 , Info fld=0x1

[RalfG]

Hi,

we have an Overland Neo Lib with 3 LTO4 HP drives. We have some trouble with the FC LTO4 drive. We saw different errors during tha last months (the other 2 drives are SCSI).

Yesterday I received a warning about a failed backup (we use bacula for backups).



12-Jun 19:13 VU0EA003-sd JobId 12786: Despooling elapsed time = 00:07:29, Transfer rate = 71.74 M bytes/second
12-Jun 19:13 VU0EA003-sd JobId 12786: Spooling data again ...
12-Jun 19:18 VU0EA003-sd JobId 12786: Job write elapsed time = 81:46:14, Transfer rate = 32.00 M bytes/second
12-Jun 19:18 VU0EA003-sd JobId 12786: Committing spooled data to Volume "A00159L4". Despooling 23,531,936,726 bytes ...
12-Jun 19:28 VU0EA003-sd JobId 12786: Fatal error: Unable to write EOF. ERR=dev.c:1681 ioctl MTWEOF error on "ULTRIUM-TD4-D3" (/dev/ULTRIUM-TD4-D3). ERR=Eingabe-/Ausgabefehler.

12-Jun 19:34 VU0EA003-sd JobId 12786: Error: Error writing final EOF to tape. This Volume may not be readable.
dev.c:1681 ioctl MTWEOF error on "ULTRIUM-TD4-D3" (/dev/ULTRIUM-TD4-D3). ERR=Eingabe-/Ausgabefehler.
12-Jun 19:34 VU0EA003-sd JobId 12786: Fatal error: spool.c:296 Fatal append error on device "ULTRIUM-TD4-D3" (/dev/ULTRIUM-TD4-D3): ERR=dev.c:1681 ioctl MTWEOF error on "ULTRIUM-TD4-D3" (/dev/ULTRIUM-TD4-D3). ERR=Eingabe-/Ausgabefehler.



The backup server shows some errors in the kernel log.


Jun 12 19:25:29 VU0EA003 kernel: [635192.318487] st2: Sense Key : Medium Error [deferred]
Jun 12 19:25:29 VU0EA003 kernel: [635192.318487] Info fld=0xfc00
Jun 12 19:25:29 VU0EA003 kernel: [635192.318487] st2: Add. Sense: Write error
Jun 12 19:28:37 VU0EA003 kernel: [635381.433471] st2: Sense Key : Medium Error [current]
Jun 12 19:28:37 VU0EA003 kernel: [635381.433471] Info fld=0x1
Jun 12 19:28:37 VU0EA003 kernel: [635381.433471] st2: Add. Sense: Write error
Jun 12 19:34:38 VU0EA003 kernel: [635742.177780] st2: Sense Key : Medium Error [current]
Jun 12 19:34:38 VU0EA003 kernel: [635742.177780] st2: Add. Sense: Recorded entity not found


All this happend after backing up ~14 TB of data. It was the end of the backup job when bacula tried to write the final EOF marker.

We are using Maxell tapes and the FC drive was already swapped 2 times. I'm a bit at a loss right now.


Drive: HP Ultrium 4-SCSI H43H
FC HBA: QLogic HPAE312A
OS: Debian Etch / AMD64


Any ideas what to do next? Any known problems with FC drives and FC HBAs?

Ralf