Trouble with AIX 3.1.5

Trouble with AIX 3.1.5

Post by Andreas Kais » Sun, 11 Aug 1991 18:19:32



I have some trouble with AIX 3.1.5:

(1) When we run some TCP/IP traffic (NFS, Telnet) via token ring and then start
dosread, the machine repeatedly crashes with a system dump  (codes 102 500
0c0).

(2) We have two disk drives connected to the machine (hdisk0 = available 400MB,
hdisk1 = available Other), but the configuration and the diagnose show three
disks (hdisk0 =  defined Other, hdisk1 = available Other, hdisk2 = available 400
MB) with hdisk0 and hdisk2 having the same  SCSI id 0. If the disgnose is booted
from diskettes, there are only hdisk0 and  1 as things should be.  Note that the
hdisk0 is marked as "defined" - not "available".
But except for (1) all is running perfectly - the machine seems  to boot from
hdisk0 and to run with hdisk2 :-).

 Obviously there is an inconsistency in the machine's configuration data.  Does
someone know now to fix this without wiping out and reinstalling the whole
machine?

                      Gruss, Andreas


 
 
 

Trouble with AIX 3.1.5

Post by Fred L. Johns » Tue, 13 Aug 1991 23:11:44



have some trouble with AIX 3.1.5:

Quote:> (1) When we run some TCP/IP traffic (NFS, Telnet) via token ring and
then start
> dosread, the machine repeatedly crashes with a system dump  (codes 102 500
> 0c0).

What device are you reading from?  If it is the diskette drive are you able to
read from the diskette using any other commands without crashing the machine?
I have seen problems with early level I/O planars that could cause
system crashes
like this.  The 102 500 0c0 means that there was an unexpected system
halt during
normal operation, that the halt was caused by an external interrupt
(most likely
a memory bus error of some sort) and that a dump was taken successfully.
 You can
send this dump into IBM support if you want.  It is probably the most useful
piece of information that we can start with.

Quote:> (2) We have two disk drives connected to the machine (hdisk0 =
available 400MB,
> hdisk1 = available Other), but the configuration and the diagnose show three
> disks (hdisk0 =  defined Other, hdisk1 = available Other, hdisk2 =
available 400
> MB) with hdisk0 and hdisk2 having the same  SCSI id 0. If the disgnose
is booted
> from diskettes, there are only hdisk0 and  1 as things should be.
Note that the
> hdisk0 is marked as "defined" - not "available".
> But except for (1) all is running perfectly - the machine seems  to boot from
> hdisk0 and to run with hdisk2 :-).

>  Obviously there is an inconsistency in the machine's configuration
data.  Does
> someone know now to fix this without wiping out and reinstalling the whole
> machine?

Since the definition for hdisk0 is redundant you should be able to just delete
it using rmdev -d -l hdisk0.  The only way that I know of for the same disk to
be identified as two different disks (one defined and one available) is for
the disk to have had its pvid changed or deleted at some point.  The pvid is
a number written on the disk (in the first physical block) that is used to
identify the disk to the system.  Even if you change SCSI ids on a disk or
move it to a different adapter it should be recognized as the same disk if the
pvid is unchanged.

It is normal for a system booted from diskette to have different names for the
hard disks if the configuration database stored on the hard disk has ever
had a change to its disk definitions (as is true in this case).  When booting
from diskette the boot operating system does not have access to the config
database stored on the hard disk, so it just identifies the first disk it
finds as hdisk0, the next hdisk1, etc.  From a diskette boot the disks are
numbered solely based on order of discovery.  This is also true on the initial
boot from hard disk but as the disk based database is modified for whatever
reason (such as disks being moved, adapters added, etc) the number ordering
may no longer match the order of discovery.

- Fred

My remarks and opinions are mine alone...
 ____________________________________________________________________________
|                                  |                                         |


| AIX BOS Field Quality            | vnet: FJOHNSON at AUSVMQ                |
| 11400 Burnet Road,  903/9320     | phone: (512) 828-2682                   |
| Austin, TX  78758-3493           | tie line: 678-2682                      |
|__________________________________|_________________________________________|

 
 
 

Trouble with AIX 3.1.5

Post by mo.. » Fri, 16 Aug 1991 00:17:19



Quote:>I have some trouble with AIX 3.1.5:

>(1) When we run some TCP/IP traffic (NFS, Telnet) via token ring and then start
>dosread, the machine repeatedly crashes with a system dump  (codes 102 500
>0c0).

This is almost surely apar ix19450.  The problem results from using the
diskette drive when there is token ring activity on the machine.  Call
IBM defect support and ask for the fix.

Quote:

>                      Gruss, Andreas

--



 
 
 

1. XMKMF trouble and AIX 4.1.1

Hi there

I have some trouble with xmkmf and AIX 4.1.1
It seems that xmkmf has some bugs. I cannot generate the Makefile from
the Imakefile... Is there a patch or bugfixes ???
The target machine is a 40P with CDE running.

                                Thanks in advance
                                  Jan Solca

2. Linux connecting to MSN

3. serial troubles with AIX 1.2.0 (PC)

4. F77 for AIX

5. sendmail troubles at AIX 4.3.3

6. Where can I get X server for Number 9 Imagine 128?

7. trouble with AIX dhcp client / Cisco dhcp server

8. MALICIOUS LATE NIGHT CALLS FROM LEE CHISNALL!!!!!!

9. Ethernet trouble with AIX

10. Trouble wt AIX as Linux NFS client

11. Trouble with AIX-based LAN. 1 machine is *very* slow!

12. More@!!.. Panasonic/SB-CD troubles, trouble, trouble...

13. Trouble, trouble, trouble - SCSI ?