SCO Openserver 6 process hang during file system access - SCO

This is a discussion on SCO Openserver 6 process hang during file system access - SCO ; Hy, I have some trouble with a new installation of SCO Openserver 6.0: Detail about machine: HP Proliant DL380 G4 (certified from vendor for SCO OS6) 2 cpu 3,40Ghz 2 hard disk 72gb (mirror, work hard disk) 1 hard disk ...

+ Reply to Thread
Results 1 to 4 of 4

Thread: SCO Openserver 6 process hang during file system access

  1. SCO Openserver 6 process hang during file system access

    Hy,
    I have some trouble with a new installation of SCO Openserver 6.0:
    Detail about machine:
    HP Proliant DL380 G4 (certified from vendor for SCO OS6)
    2 cpu 3,40Ghz
    2 hard disk 72gb (mirror, work hard disk)
    1 hard disk 72 gb (single drive, backup only)
    4 GB Ram
    Sco OS 6.0 + mp1 +mp2 +oss706b+oss702b+oss701a
    HP EFS support pack 7.74
    Latest BIOS/HD Firmware/ecc.ecc.
    85 Users

    System seems to running fine, but, apparently in a random way, for example
    accessing the backup disk with commands like "df - kv", the process hang
    without warning. After that, access to the backup filesystem is not more
    possible. There are not traces of events in syslogd and adm/messages. The
    customers of the system are not influenced from the problem, but in order
    re-enable the access to the backup filesystem it is necessary to execute a
    reboot.
    Anyone have an idea, or have observed a similar issue ?
    There is a way to trace that kind of problem, like "live debug" of the
    system ?
    (Keeep in mind that is a production enviroment, every experiment need to be
    a bit "secure").
    The only "strange" setup i have noticed is that the Operating System
    (installation is made from other technician), have the first hard disk
    formatted WITHOUT using vxfs (default of installation program in case of
    fresh install).
    But we need to use large file support on the spare drive for backup purpose,
    so ve have formatted this with vxfs.

    +-------------------+------------+--------+---+-------------+------------+
    | Name | Type | New FS | # | First Block | Last Block |
    +-------------------+------------+--------+---+-------------+------------+
    | boot | EAFS | no | 0 | 8032| 48990|
    | dump | NON FS | no | 1 | 48991| 2782045|
    | root | HTFS | no | 2 | 2782046| 10974044|
    | (label = /usr3) | HTFS | no | 3 | 10974045| 15837019|
    | (label = /u) | HTFS | no | 4 | 15837020| 16861018|
    | | NOT USED | no | 5 | -| -|
    | | NOT USED | no | 6 | -| -|
    | hd0a | WHOLE DISK | no | 7 | 0| 71118606|
    | (label = /usr1) | HTFS | no | 12| 16861019| 55773017|
    | | NON FS | no | 13| 0| 8015|
    | (label = /usr2) | HTFS | no | 14| 55773018| 71118600|
    +-------------------+------------+--------+---+-------------+------------+
    71118607 1K blocks for divisions, 8016 1K blocks reserved for the system

    +-------------------+------------+--------+---+-------------+------------+
    | Name | Type | New FS | # | First Block | Last Block |
    +-------------------+------------+--------+---+-------------+------------+
    | usr4 | vxfs | no | 0 | 0| 71118606|
    | d00101 | NOT USED | no | 1 | -| -|
    | d00102 | NOT USED | no | 2 | -| -|
    | d00103 | NOT USED | no | 3 | -| -|
    | d00104 | NOT USED | no | 4 | -| -|
    | d00105 | NOT USED | no | 5 | -| -|
    | d00106 | NOT USED | no | 6 | -| -|
    | d00107all | WHOLE DISK | no | 7 | 0| 71126622|
    +-------------------+------------+--------+---+-------------+------------+
    71118607 1K blocks for divisions, 8016 1K blocks reserved for the system

    When i observed this "hang" for the first time, i have tried to remove a
    tape drive (DAT72) that is originally installed on the same SCSI channel of
    second hard drive,
    (I dont like using a SCSI Raid controller with hard disk AND tape drive, for
    a geneal rule), but without success.

    Any helps is welcome,

    Mirko



  2. Re: SCO Openserver 6 process hang during file system access

    FWIW I have an almost identical configuration.

    You could try looking at the HP EFS diagnostic menu:

    /usr/bin/compaq/diags/ciss/ciss_menu

    But how is this "backup only" drive used? I'm unclear why a mirror
    needs a backup. I suspect some mischief in however copying is done from
    the mirror to this drive.

    --Ray Robert

    On Jan 23, 2:55 pm, "news.tele2.it" wrote:
    > Hy,
    > I have some trouble with a new installation of SCO Openserver 6.0:
    > Detail about machine:
    > HP Proliant DL380 G4 (certified from vendor for SCO OS6)
    > 2 cpu 3,40Ghz
    > 2 hard disk 72gb (mirror, work hard disk)
    > 1 hard disk 72 gb (single drive, backup only)
    > 4 GB Ram
    > Sco OS 6.0 + mp1 +mp2 +oss706b+oss702b+oss701a
    > HP EFS support pack 7.74
    > Latest BIOS/HD Firmware/ecc.ecc.
    > 85 Users
    >
    > System seems to running fine, but, apparently in a random way, for example
    > accessing the backup disk with commands like "df - kv", the process hang
    > without warning. After that, access to the backup filesystem is not more
    > possible. There are not traces of events in syslogd and adm/messages. The
    > customers of the system are not influenced from the problem, but in order
    > re-enable the access to the backup filesystem it is necessary to execute a
    > reboot.
    > Anyone have an idea, or have observed a similar issue ?
    > There is a way to trace that kind of problem, like "live debug" of the
    > system ?
    > (Keeep in mind that is a production enviroment, every experiment need to be
    > a bit "secure").
    > The only "strange" setup i have noticed is that the Operating System
    > (installation is made from other technician), have the first hard disk
    > formatted WITHOUT using vxfs (default of installation program in case of
    > fresh install).
    > But we need to use large file support on the spare drive for backup purpose,
    > so ve have formatted this with vxfs.
    >
    > +-------------------+------------+--------+---+-------------+------------+
    > | Name | Type | New FS | # | First Block | Last Block |
    > +-------------------+------------+--------+---+-------------+------------+
    > | boot | EAFS | no | 0 | 8032| 48990|
    > | dump | NON FS | no | 1 | 48991| 2782045|
    > | root | HTFS | no | 2 | 2782046| 10974044|
    > | (label = /usr3) | HTFS | no | 3 | 10974045| 15837019|
    > | (label = /u) | HTFS | no | 4 | 15837020| 16861018|
    > | | NOT USED | no | 5 | -| -|
    > | | NOT USED | no | 6 | -| -|
    > | hd0a | WHOLE DISK | no | 7 | 0| 71118606|
    > | (label = /usr1) | HTFS | no | 12| 16861019| 55773017|
    > | | NON FS | no | 13| 0| 8015|
    > | (label = /usr2) | HTFS | no | 14| 55773018| 71118600|
    > +-------------------+------------+--------+---+-------------+------------+
    > 71118607 1K blocks for divisions, 8016 1K blocks reserved for the system
    >
    > +-------------------+------------+--------+---+-------------+------------+
    > | Name | Type | New FS | # | First Block | Last Block |
    > +-------------------+------------+--------+---+-------------+------------+
    > | usr4 | vxfs | no | 0 | 0| 71118606|
    > | d00101 | NOT USED | no | 1 | -| -|
    > | d00102 | NOT USED | no | 2 | -| -|
    > | d00103 | NOT USED | no | 3 | -| -|
    > | d00104 | NOT USED | no | 4 | -| -|
    > | d00105 | NOT USED | no | 5 | -| -|
    > | d00106 | NOT USED | no | 6 | -| -|
    > | d00107all | WHOLE DISK | no | 7 | 0| 71126622|
    > +-------------------+------------+--------+---+-------------+------------+
    > 71118607 1K blocks for divisions, 8016 1K blocks reserved for the system
    >
    > When i observed this "hang" for the first time, i have tried to remove a
    > tape drive (DAT72) that is originally installed on the same SCSI channel of
    > second hard drive,
    > (I dont like using a SCSI Raid controller with hard disk AND tape drive, for
    > a geneal rule), but without success.
    >
    > Any helps is welcome,
    >
    > Mirko



  3. Re: SCO Openserver 6 process hang during file system access

    Simple a cron job do a cpio-backup (of the data present on the first disk)
    on this disk, and another job follow to save this file on a remote tape
    library.
    Disk-to-Disk cpio is very fast, application running on this system is an
    ISAM database, and all-user need to be offline when a backup is made.

    You have a dual-processor system ?
    Do you have the first disk formatted with VXFS ?

    Do you know is thare is an utility like "strace" in OS6 for live-debug a
    running process ?

    Thanks,

    Mirko


    > FWIW I have an almost identical configuration.
    >
    > You could try looking at the HP EFS diagnostic menu:
    >
    > /usr/bin/compaq/diags/ciss/ciss_menu
    >
    > But how is this "backup only" drive used? I'm unclear why a mirror
    > needs a backup. I suspect some mischief in however copying is done from
    > the mirror to this drive.
    >
    > --Ray Robert
    >
    > On Jan 23, 2:55 pm, "news.tele2.it" wrote:
    >> Hy,
    >> I have some trouble with a new installation of SCO Openserver 6.0:
    >> Detail about machine:
    >> HP Proliant DL380 G4 (certified from vendor for SCO OS6)
    >> 2 cpu 3,40Ghz
    >> 2 hard disk 72gb (mirror, work hard disk)
    >> 1 hard disk 72 gb (single drive, backup only)
    >> 4 GB Ram
    >> Sco OS 6.0 + mp1 +mp2 +oss706b+oss702b+oss701a
    >> HP EFS support pack 7.74
    >> Latest BIOS/HD Firmware/ecc.ecc.
    >> 85 Users
    >>
    >> System seems to running fine, but, apparently in a random way, for
    >> example
    >> accessing the backup disk with commands like "df - kv", the process hang
    >> without warning. After that, access to the backup filesystem is not more
    >> possible. There are not traces of events in syslogd and adm/messages. The
    >> customers of the system are not influenced from the problem, but in order
    >> re-enable the access to the backup filesystem it is necessary to execute
    >> a
    >> reboot.
    >> Anyone have an idea, or have observed a similar issue ?
    >> There is a way to trace that kind of problem, like "live debug" of the
    >> system ?
    >> (Keeep in mind that is a production enviroment, every experiment need to
    >> be
    >> a bit "secure").
    >> The only "strange" setup i have noticed is that the Operating System
    >> (installation is made from other technician), have the first hard disk
    >> formatted WITHOUT using vxfs (default of installation program in case of
    >> fresh install).
    >> But we need to use large file support on the spare drive for backup
    >> purpose,
    >> so ve have formatted this with vxfs.
    >>
    >> +-------------------+------------+--------+---+-------------+------------+
    >> | Name | Type | New FS | # | First Block | Last Block
    >> |
    >> +-------------------+------------+--------+---+-------------+------------+
    >> | boot | EAFS | no | 0 | 8032|
    >> 48990|
    >> | dump | NON FS | no | 1 | 48991|
    >> 2782045|
    >> | root | HTFS | no | 2 | 2782046|
    >> 10974044|
    >> | (label = /usr3) | HTFS | no | 3 | 10974045|
    >> 15837019|
    >> | (label = /u) | HTFS | no | 4 | 15837020|
    >> 16861018|
    >> | | NOT USED | no | 5
    >> -| -|
    >> | | NOT USED | no | 6
    >> -| -|
    >> | hd0a | WHOLE DISK | no | 7 | 0|
    >> 71118606|
    >> | (label = /usr1) | HTFS | no | 12| 16861019|
    >> 55773017|
    >> | | NON FS | no | 13| 0|
    >> 8015|
    >> | (label = /usr2) | HTFS | no | 14| 55773018|
    >> 71118600|
    >> +-------------------+------------+--------+---+-------------+------------+
    >> 71118607 1K blocks for divisions, 8016 1K blocks reserved for the system
    >>
    >> +-------------------+------------+--------+---+-------------+------------+
    >> | Name | Type | New FS | # | First Block | Last Block
    >> |
    >> +-------------------+------------+--------+---+-------------+------------+
    >> | usr4 | vxfs | no | 0 | 0|
    >> 71118606|
    >> | d00101 | NOT USED | no | 1
    >> -| -|
    >> | d00102 | NOT USED | no | 2
    >> -| -|
    >> | d00103 | NOT USED | no | 3
    >> -| -|
    >> | d00104 | NOT USED | no | 4
    >> -| -|
    >> | d00105 | NOT USED | no | 5
    >> -| -|
    >> | d00106 | NOT USED | no | 6
    >> -| -|
    >> | d00107all | WHOLE DISK | no | 7 | 0|
    >> 71126622|
    >> +-------------------+------------+--------+---+-------------+------------+
    >> 71118607 1K blocks for divisions, 8016 1K blocks reserved for the system
    >>
    >> When i observed this "hang" for the first time, i have tried to remove a
    >> tape drive (DAT72) that is originally installed on the same SCSI channel
    >> of
    >> second hard drive,
    >> (I dont like using a SCSI Raid controller with hard disk AND tape drive,
    >> for
    >> a geneal rule), but without success.
    >>
    >> Any helps is welcome,
    >>
    >> Mirko

    >




  4. Re: SCO Openserver 6 process hang during file system access

    news.tele2.it typed (on Wed, Jan 24, 2007 at 08:58:29PM +0100):
    | Do you know is thare is an utility like "strace" in OS6 for live-debug a
    | running process ?

    /bin/truss ?

    --
    JP
    ==> http://www.frappr.com/cusm <==

+ Reply to Thread