From: Goetz Schuchart (goetzs@uranus.central.de)
Date: 08/01/93


From: goetzs@uranus.central.de (Goetz Schuchart)
Subject: ADAPTEC: SCSI host x timed out - ... HILFE !
Date: Sun, 1 Aug 1993 22:27:58 GMT


        I posted this message one week ago - but I did not get
        very much help. What I got were many messages from people
        who have the same problem:

        Many people who use more than one SCSI drive get these
        "SCSI host 0 timed out - Aborting command"-errors mostly
        caused by heavy disk activity...

        I would be pleased if you could forward this message
        to the author of the Adaptec drivers for Linux and
        for any help I can get with this strange problem (had
        crashed my filesystem three times until now... what
        meant a lot of work for me...)

  --- problem description ---

        Configuration:

        Hardware:

        486DX50-EISA, 8 MB ram, Adaptec 1742 with ST3283N and
        ST1096N, AST Fourport, IDE-Controller with Conner and
        ST3283A

        Software:

        Kernel 0.99-pl9

        When having really high diskactivity on both scsi-hds
        _and_ on one IDE-hd _at the same time_ I get a kernel
        panic:

        Unable to handle kernel paging request at address c00000cc
        Oops: 0000
        EIP: 0008:00061P8E
        EFLAGS: 00010202
        fs: 0017
        base: C000000000 Limit: 00200000
        Pid 0, process nr: 0
        83 bdcc 00 00 00 00 74 14 68
        task[0] (swapper) Killed: unable to recover
        Kernel panic: Trying to free up swapper memory space
        In swapper task - not syncing
        SCSI host 0 timed out - aborting command
        SCSI host 0 timed out - aborting command

        Now I want to give an example when this error
        occurs every time:

        / is Conner IDE
        /usr/spool is Seagate SCSI (ST3283N)
        /mnt2 is Seagate SCSI (ST1096N)
        swap is Seagate IDE (ST3283A)

        cd /u
        find . -xdev | cpio -paduv /mnt2 &
        cd /usr/spool
        cp Log.o x.x &
        cp Log.oo x.y &
        sync

        All tasks are in background and work at the same time !
        Log.o and Log.oo are big files (over 1 MByte)
        What I do is backing up the /u tree from IDE-disk to
        SCSI-disk and copying two big files on scsi-Disk
        all at the same time. This means high disk activity
        on three disks.

        Every time I do this (as described above) I get this
        error...

        With kind regards,
                Goetz Schuchart

        *harghl* - ich habe Kopfschmerzen...

-- 
      Goetz Schuchart - goetzs@uranus.central.de - postmaster@central.de
    Koenigsberger Str.13  *  34270 Schauenburg  *  -FRG-  *  ++49 5601 1243

-- Goetz Schuchart - goetzs@uranus.central.de - postmaster@central.de Koenigsberger Str.13 * 34270 Schauenburg * -FRG- * ++49 5601 1243