Messages in this thread | | | From | "Stephen Davies" <> | Subject | SCSI problems (fup to backup to Exabyte post) | Date | Mon, 15 Apr 96 22:42:37 +1030 |
| |
G'day all.
I have now updated to 1.3.88 and my backup to Exabyte problems seem to have escalated.
As well as hanging my SCSI controller, backups from SCSI disk to SCSI tape now consistently cause SCSI disk file system corruption.
So far today, I have tried two separate tape units (one straight from the factory), two DIP switch settings and two different source disks for testing.
All except an IDE to SCSI tape have resulted in a hung Adaptec 1542 controller and a corrupted SCSI-based file system.
The file system errors do not seem to have a pattern: I have seen both "Duplicate or bad blocks in use" and "Bad inode for . in directory inode x" errors. (And incidentally, I can no longer unmount the SCSI disk to run e2fsck: I have to edit fstab and reboot before I can run e2fsck to (thankfully) repair at least most of the damage.
My messages log is the same as last time (another copy attached below) but now I can also see two copies of the following:
"scsi:aborting command due to timeout:pid 22313,scsi 0,channel 0, id 1, lun 0 0x08 1c 13 47 1c 00"
ID 1 is my SCSI disk. (the Exabyte is ID 6 and the 1542 is ID 7)
I have also tried high volume transfers from SCSI disk to the same disk but this does not fail: it seems to be only SCSI disk to tape transfers that cause me grief.
I had a signifiant number of "me too" responses to my last post but only two positive suggestions (to do with Exabyte DIP switch settings which I am trying to follow up with Exabyte) but no comments re Linux SCSI support code. I cannot help but believe that the problem _must_ be at least partially a Linux problem as it has only arisen with the 1.3.x kernel series and seems to be getting worse with each new release.
Cheers, Stephen.
Apr 15 17:57:56 mustang kernel: Configuring Adaptec (SCSI-ID 7) at IO:330, IRQ 1 1, DMA priority 5 Apr 15 17:57:56 mustang kernel: scsi0 : Adaptec 1542 Apr 15 17:57:56 mustang kernel: scsi : 1 host. Apr 15 17:57:57 mustang kernel: Vendor: IBM Model: DPES-31080 Rev : S31Q Apr 15 17:57:57 mustang kernel: Type: Direct-Access ANS I SCSI revision: 02 Apr 15 17:57:57 mustang kernel: Detected scsi disk sda at scsi0, channel 0, id 1 , lun 0 Apr 15 17:57:57 mustang kernel: Vendor: EXABYTE Model: EXB-8200 Rev : 2618 Apr 15 17:57:57 mustang kernel: Type: Sequential-Access ANS I SCSI revision: 01 Apr 15 17:57:57 mustang kernel: scsi : detected 1 SCSI disk total. Apr 15 17:57:57 mustang kernel: SCSI device sda: hdwr sector= 512 bytes. Sectors = 2118144 [1034 MB] [1.0 GB] Apr 15 17:57:57 mustang kernel: wd.c:v1.10 9/23/94 Donald Becker (becker@cesdis. gsfc.nasa.gov) Apr 15 17:57:57 mustang kernel: eth0: WD80x3 at 0x300, 00 00 C0 47 39 5C WD8013 , IRQ 10, shared memory at 0xd8000-0xdbfff. Apr 15 17:57:57 mustang kernel: Partition check: Apr 15 17:57:57 mustang kernel: sda: sda1 Apr 15 17:57:57 mustang kernel: hda: hda1 hda2 hda3 Apr 15 17:57:57 mustang kernel: hdb: hdb1 Apr 15 17:57:58 mustang kernel: VFS: Mounted root (ext2 filesystem) readonly. Apr 15 17:57:58 mustang kernel: Adding Swap: 12396k swap-space Apr 15 17:58:19 mustang kernel: scsi : aborting command due to timeout : pid 467 21, scsi0, channel 0, id 1, lun 0 0x08 1d 6e ff 10 00 Apr 15 17:58:19 mustang kernel: SCSI host 0 abort (pid 46721) timed out - resett ing Apr 15 17:58:20 mustang kernel: SCSI bus is being reset for host 0. Apr 15 17:58:20 mustang kernel: HACC aha1542.c: interrupt received, but no mail. Apr 15 17:58:20 mustang kernel: Sent BUS RESET to scsi host 0 Apr 15 17:58:21 mustang kernel: Sending DID_RESET for target 1 Apr 15 17:58:21 mustang last message repeated 7 times Apr 15 17:58:21 mustang kernel: aha1542_intr_handle: Unexpected interrupt Apr 15 17:58:21 mustang kernel: tarstat=0, hastat=0 idlun=28 ccb#=3 Apr 15 17:58:21 mustang kernel: aha1542_intr_handle: Unexpected interrupt Apr 15 17:58:21 mustang kernel: tarstat=0, hastat=0 idlun=28 ccb#=4 Apr 15 17:58:29 mustang kernel: scsi : aborting command due to timeout : pid 467 21, scsi0, channel 0, id 1, lun 0 0x08 1d 6e ff 10 00 Apr 15 17:58:29 mustang kernel: SCSI host 0 abort (pid 46721) timed out - resett ing Apr 15 17:58:29 mustang kernel: SCSI bus is being reset for host 0. Apr 15 17:58:29 mustang kernel: HACC aha1542.c: interrupt received, but no mail. Apr 15 17:58:29 mustang kernel: Sent BUS RESET to scsi host 0 Apr 15 17:58:29 mustang kernel: Sending DID_RESET for target 1 Apr 15 17:58:29 mustang last message repeated 6 times Apr 15 17:58:29 mustang kernel: aha1542_intr_handle: Unexpected interrupt Apr 15 17:58:29 mustang kernel: tarstat=2, hastat=0 idlun=28 ccb#=3 Apr 15 17:58:29 mustang kernel: aha1542_intr_handle: Unexpected interrupt Apr 15 17:58:29 mustang kernel: tarstat=0, hastat=0 idlun=28 ccb#=4 Apr 15 17:58:29 mustang kernel: aha1542_intr_handle: Unexpected interrupt Apr 15 17:58:29 mustang kernel: tarstat=0, hastat=0 idlun=28 ccb#=5 Apr 15 17:58:29 mustang kernel: aha1542_intr_handle: Unexpected interrupt Apr 15 17:58:29 mustang kernel: tarstat=0, hastat=0 idlun=28 ccb#=6 Apr 15 17:58:29 mustang kernel: aha1542_intr_handle: Unexpected interrupt Apr 15 17:58:29 mustang kernel: tarstat=0, hastat=0 idlun=28 ccb#=7
======================================================================== Stephen Davies Consulting scldad@sdc.com.au Adelaide, South Australia. Voice: 61-8-2728863 Computing & Network solutions. Fax : 61-8-2741015
| |