home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!Germany.EU.net!ira.uka.de!uka!i11s10!wolpers
- From: wolpers@i11s10.ira.uka.de (Andreas Wolpers)
- Newsgroups: comp.sys.sun.hardware
- Subject: Problems with Fujitsu 2654SA disk and SunOs 4.1.2
- Date: 17 Dec 1992 22:04:54 GMT
- Organization: University of Karlsruhe, FRG
- Lines: 93
- Distribution: world
- Message-ID: <1gqti6INNjuj@iraul1.ira.uka.de>
- NNTP-Posting-Host: i11s10.ira.uka.de
- Keywords: dump, large disks
-
- Hi folks,
-
- last friday we installed two Fujitsu M2654SA disks (2 GByte) on two
- Sparstation 2 running SunOS 4.1.2.
-
- They were delivered together, one formatted, one unformatted.
- Well, I fomatted the second one with parameters taken from the first
- disks label, and 3600 rpm (since the label didn't specify this value
- and I couldn't find any information in the accompanying "literature".
-
- Initally, both were filles with 450-490 MB of data (home directories :-( ).
- The increase in disk usage in the following days was minimal (both had
- less than 500MB when the nightmare started)
-
- Both disks worked fine 'til Tuesday, 2am. That's when our weekly
- level 0 dump begins. The first disk (which came formatted) survived
- this event, the second failed in the middle of the process, garbled
- beyond recognition. I called the vendor and since they told me that
- this was a 5400 rpm disk, I thought it was my fault.
-
- Until this afternoon.
-
- We did another level 0 dump last night, and in the corresponding log file
- I found a message
-
- DUMP: (This should not happen)bread from /dev/rsd2c [block 34640]: count=8192,got=-1
-
-
- which made me a bit nervous after the crash on tuesday morning. A look at
- the logs of the level5 dumps in between revealed messages like
-
- DUMP: corrupted directory, inumber 342411
-
- I unmounted the file system (The entire disks were used as 2GB partitions)
- and fsck'ed it: no errors. I dumped it to /dev/null and got lots of messages
- like the ones above. Another fsck found no errors.
-
- Finally, I tried to tar the disk's contents to a tape. FUBAR!!!!!!
-
- /var/adm/messages contains one disk related message from Dec. 15, and lots of
- messages from this afternoon:
-
- Dec 15 02:11:47 ardbeg vmunix: esp0: Current command timeout for Target 2 Lun 0
- Dec 15 02:11:47 ardbeg vmunix: esp0: State=UNKNOWN (0x1a), Last State=RESEL (0x1b)
- Dec 15 02:11:47 ardbeg vmunix: esp0: Cmd dump for Target 2 Lun 0:
- Dec 15 02:11:47 ardbeg vmunix: esp0: cdb=[ 0x8 0x5 0x7f 0xd0 0x6 0x0 0x0 0x0 0x0 0x0 ]
- Dec 15 02:11:47 ardbeg vmunix: sd2: SCSI transport failed: reason 'reset': retrying command
- Dec 15 02:11:47 ardbeg vmunix: sd1: SCSI transport failed: reason 'reset': retrying command
- Dec 15 02:11:47 ardbeg vmunix: esp0: Target 1 now Synchronous at 4.0 mb/s max transmit rate
-
-
-
- Dec 17 02:14:15 ardbeg vmunix: esp0: Current command timeout for Target 2 Lun 0
- Dec 17 02:14:15 ardbeg vmunix: esp0: State=UNKNOWN (0x1a), Last State=RESEL (0x1b)
- Dec 17 02:14:15 ardbeg vmunix: esp0: Cmd dump for Target 2 Lun 0:
- Dec 17 02:14:15 ardbeg vmunix: esp0: cdb=[ 0x8 0x0 0x87 0x32 0xc 0x0 0x0 0x0 0x0 0x0 ]
- Dec 17 02:14:15 ardbeg vmunix: sd2: SCSI transport failed: reason 'reset': retrying command
- Dec 17 02:14:18 ardbeg vmunix: esp0: Target 1 now Synchronous at 4.0 mb/s max transmit rate
- Dec 17 02:14:18 ardbeg vmunix: esp0: Target 1 now Synchronous at 4.0 mb/s max transmit rate
- Dec 17 02:14:18 ardbeg vmunix: sd2c: Error for command 'read'
- Dec 17 02:14:18 ardbeg vmunix: sd2c: Error Level: Fatal
- Dec 17 02:14:18 ardbeg vmunix: sd2c: Block 34640, Absolute Block: 34640
- Dec 17 02:14:18 ardbeg vmunix: sd2c: Sense Key: Hardware Error
- Dec 17 02:14:18 ardbeg vmunix: sd2c: Vendor 'FUJITSU' error code: 0x15
-
- Dec 17 15:40:57 ardbeg vmunix: esp0: Disconnected command timeout for Target 2 Lun 0
- Dec 17 15:40:57 ardbeg vmunix: sd2: SCSI transport failed: reason 'timeout': retrying command
-
-
- Dec 17 16:34:08 ardbeg vmunix: esp0: Current command timeout for Target 2 Lun 0
- Dec 17 16:34:08 ardbeg vmunix: esp0: State=UNKNOWN (0x1a), Last State=RESEL (0x1b)
- Dec 17 16:34:08 ardbeg vmunix: esp0: Cmd dump for Target 2 Lun 0:
- Dec 17 16:34:08 ardbeg vmunix: esp0: cdb=[ 0x28 0x0 0x0 0x30 0xcd 0x0 0x0 0x0 0x2 0x0 ]
- Dec 17 16:34:08 ardbeg vmunix: sd2: SCSI transport failed: reason 'reset': retrying command
- Dec 17 16:34:08 ardbeg vmunix: sd1: SCSI transport failed: reason 'reset': retrying command
- Dec 17 16:34:08 ardbeg vmunix: esp0: Target 1 now Synchronous at 4.0 mb/s max transmit rate
-
- Dec 17 17:02:50 ardbeg vmunix: bad block 100221746, <3>/mnt: bad block
-
- Dec 17 21:45:25 ardbeg vmunix: sd2 at esp0 target 2 lun 0
- Dec 17 21:45:25 ardbeg vmunix: sd2: <Fujitsu-M2654SA cyl 2195 alt 2 hd 21 sec 87>
-
- Any idea what's going on?
- Oh,btw, ardbeg has no internal drives, an external 520 MB Fujitsu 3.5 inch
- and the above-mentioned 2 GB drive. about 2.5 meters cable externally.
-
-
-
- --
- Andreas Wolpers University of Karlsruhe
- e-mail: wolpers@ira.uka.de Institute for Logic, Complexity
- phone: 49-721-608-3977 and Deductive Systems
- fax: 49-721-697760 Karlsruhe, Germany
-