Hi. I want to setup an HP SureStore T20 streamer on a box with a newly installed RH9 Linux. The computer is Intel P3 800 MUh with 512 MB RAM. It's a small-office server used to distribute files to the intranet and providing internet capabilities (squid, apache, postfix, fetchmail, ...). I'm using Travan tapes with 10/20 GB (compressed) size. The streamer is connected to an SCSI controller DawiControl DC-2974 PCI (scsi0: Tekram DC390/AM53C974 V2.0f 2000-12-20). There is a second one (scsi1: 3ware Storage Controller) for a 3ware RAID. The system contains an IDE CD/DVD drive, two 3Com PCI 3c905C Tornado Vers LK1.1.18-ac cards, and several on-board devices (audio, usb). The old OS was an SuSE Linux 7.0 and employees meant that the streamer worked during this time. I can't imagine why it makes problems now :-( The computer crashes if I'm storing big (several GB) directories. I'm starting the backup and after some hours the system hangs and nothing is able to reanimate the OS. I tested with these software packages: arkeia light, amanda and mt. Simple tests using tar and small files can be written and reread without any problems. A "mt rewind ; tar cf /dev/tape /data" allows me to reproduce the bug. After some time nothing works anymore, so I think it's not a client problem. Tested kernels were the standard RH9 kernel and an update I got from the RHN (version 2.4.20-20.9). The kernel log contains a lot of lines like the following: Nov 6 03:31:02 linux kernel: DC390: Pointer restored. Total -12736512, Bus bf7d2800 Nov 6 03:31:04 linux kernel: DC390: Pointer restored. Total -12703744, Bus bf7da800 Nov 6 03:31:04 linux kernel: DC390: Pointer restored. Total -12703744, Bus bf7da800 Nov 6 03:31:06 linux kernel: DC390: Pointer restored. Total -12670976, Bus bf7e2800 Nov 6 03:31:06 linux kernel: DC390: Pointer restored. Total -12670976, Bus bf7e2800 Nov 6 03:31:08 linux kernel: DC390: Pointer restored. Total -12638208, Bus bf7ea800 Nov 6 03:31:08 linux kernel: DC390: Pointer restored. Total -12638208, Bus bf7ea800 Nov 6 03:31:10 linux kernel: DC390: Pointer restored. Total -12605440, Bus bf7f2800 Nov 6 03:31:10 linux kernel: DC390: Pointer restored. Total -12605440, Bus bf7f2800 Nov 6 03:31:12 linux kernel: DC390: Pointer restored. Total -12572672, Bus bf7fa800 after some time I see: Nov 6 20:04:06 linux kernel: scsi : aborting command due to timeout: pid 450046, scsi0, channel 0, id 4, lun 0 Prevent/Allow Medium Removal 00 00 00 01 00 Nov 6 20:04:06 linux kernel: DC390: Abort command (pid 450046, Device 04-00) Nov 6 20:04:06 linux kernel: DC390: SRB: Xferred 00000000, Remain 00000000, State 00000040, Phase 05 Nov 6 20:04:06 linux kernel: DC390: AdpaterStatus: 00, SRB Status 00 Nov 6 20:04:06 linux kernel: DC390: Status of last IRQ (DMA/SC/Int/IRQ): 2080c420 Nov 6 20:04:06 linux kernel: DC390: Register dump: SCSI block: Nov 6 20:04:06 linux kernel: DC390: XferCnt Cmd Stat IntS IRQS FFIS Ctl1 Ctl2 Ctl3 Ctl4 Nov 6 20:04:06 linux kernel: DC390: 000028 42 02 c3 00 60 17 48 08 84 Nov 6 20:04:06 linux kernel: DC390: Register dump: DMA engine: Nov 6 20:04:06 linux kernel: DC390: Cmd STrCnt SBusA WrkBC WrkAC Stat SBusCtrl Nov 6 20:04:06 linux kernel: DC390: 00 00000040 1daa1710 00000024 1daa172c 00 03184500 Nov 6 20:04:06 linux kernel: DC390: Register dump: PCI Status: 0200 Nov 6 20:04:06 linux kernel: DC390: In case of driver trouble read linux/drivers/scsi/README.tmscsim Nov 6 20:04:06 linux kernel: DC390: Abort current command (pid 450046, SRB dfe90144) Nov 6 20:04:06 linux kernel: DC390: Aborted pid 450046 with status 3 Nov 6 20:04:06 linux kernel: SCSI host 0 abort (pid 450046) timed out - resetting Nov 6 20:04:06 linux kernel: SCSI bus is being reset for host 0 channel 0. Nov 6 20:04:06 linux kernel: DC390: RESET ... done Nov 6 20:04:06 linux kernel: DC390: Illegal Operation detected (20c38418)! Nov 6 20:04:06 linux kernel: DC390: SRB: Xferred 00000000, Remain 00000000, State 00000040, Phase 05 Nov 6 20:04:06 linux kernel: DC390: AdpaterStatus: 00, SRB Status 00 Nov 6 20:04:06 linux kernel: DC390: Status of last IRQ (DMA/SC/Int/IRQ): 20c38418 ... shortly after that event everything stops. Sometimes there is also a small stack trace: Nov 6 20:29:34 linux kernel: do_IRQ: stack overflow: 956 Nov 6 20:29:34 linux kernel: d1bf4a6c 000003bc c030e530 000c3244 00000000 ddaa1600 dfe90000 c010d778 Nov 6 20:29:34 linux kernel: 000c3244 b131c277 00001fa6 00000000 ddaa1600 dfe90000 b1391cc0 ccca0068 Nov 6 20:29:34 linux kernel: cd650068 ffffff00 c024b964 00000060 00000287 00000002 e080d8ed 000c3244 Nov 6 20:29:34 linux kernel: Call Trace: [<c010d778>] call_do_IRQ [kernel] 0x5 (0xd1bf4a88)) Nov 6 20:29:34 linux kernel: [<c024b964>] __rdtsc_delay [kernel] 0x14 (0xd1bf4ab4)) Nov 6 20:29:34 linux kernel: [<e080d8ed>] scsi_dispatch_cmd [scsi_mod] 0x33d (0xd1bf4ac4)) Nov 6 20:29:34 linux kernel: [<e0816446>] scsi_request_fn [scsi_mod] 0x1d6 (0xd1bf4afc)) Nov 6 20:29:34 linux kernel: [<e087be0c>] rh_init_int_timer [usb-uhci] 0x5c (0xd1bf4b0c)) Nov 6 20:29:34 linux kernel: [<e0815818>] __scsi_insert_special [scsi_mod] 0x58 (0xd1bf4b34)) Nov 6 20:29:34 linux kernel: [<e0815898>] scsi_insert_special_req [scsi_mod] 0x28 (0xd1bf4b44)) Nov 6 20:29:34 linux kernel: [<e080daab>] scsi_do_req_R1f341175 [scsi_mod] 0xeb (0xd1bf4b58)) Nov 6 20:29:34 linux kernel: [<e082e965>] dc390_SendSRB [tmscsim] 0x85 (0xd1bf4b6c)) Nov 6 20:29:34 linux kernel: [<e080d0b0>] scsi_wait_done [scsi_mod] 0x0 (0xd1bf4b8c)) Nov 6 20:29:34 linux kernel: [<e080d97d>] scsi_wait_req_R6f82968f [scsi_mod] 0x6d (0xd1bf4ba8)) Nov 6 20:29:34 linux kernel: [<e080d0b0>] scsi_wait_done [scsi_mod] 0x0 (0xd1bf4bbc)) Nov 6 20:29:34 linux kernel: [<e0810a08>] ioctl_internal_command [scsi_mod] 0x68 (0xd1bf4be4)) Nov 6 20:29:34 linux kernel: [<e08110c1>] scsi_ioctl_R4cb00872 [scsi_mod] 0x101 (0xd1bf4c10)) Nov 6 20:29:34 linux kernel: [<e08165ce>] scsi_request_fn [scsi_mod] 0x35e (0xd1bf4c40)) Nov 6 20:29:34 linux kernel: [<e0815a3d>] scsi_queue_next_request [scsi_mod] 0x3d (0xd1bf4c78)) Nov 6 20:29:34 linux kernel: [<e080d599>] scsi_release_command_Rf1aeb218 [scsi_mod] 0x29 (0xd1bf4c90)) Nov 6 20:29:34 linux kernel: [<e080d9aa>] scsi_wait_req_R6f82968f [scsi_mod] 0x9a (0xd1bf4ca0)) Nov 6 20:29:34 linux kernel: [<e080d0b0>] scsi_wait_done [scsi_mod] 0x0 (0xd1bf4cb4)) (last block repeated multiple times) I looked into the DC390 driver and I can't find a reason for the crash. The last driver code change occured in 2000 I believe. With Google I couldn't find any relevant information. Is this error known and do others have problems with such a configuration? Is there a better driver I could use for this streamer? If yes, where? What can I do to remove the problem? Every hint and any idea is possibly helpful. Thank you for reading this long mail. Regards, Gerrit Albrecht -- redhat-list mailing list unsubscribe mailto:redhat-list-request@xxxxxxxxxx?subject=unsubscribe https://www.redhat.com/mailman/listinfo/redhat-list
