Fatal error: building package index (custom.tlib)

[antix]

As short as I can make it. This system has had a bad history with TD's. So I am remotely trying to determine the issues. This system is an HP NETSERVER 6/300 LH PRO II (64MB) running SCO 5.0.4. Running dual processors. The last time they replaced a TD, the system wouldn't recognize it so they asked me to dial-up, remove and re-add the TD. I thought this was rather curious because they said it was like for like and config the same. Well during the initial remove the kernel re-link bombs out.

Sooo I was trying to add a patch to fix a problem with the kernel rebuild. That error:
Root for this system build is /
"/var/opt/K/SCO/link/1.1.0Eb/etc/conf/pack.d/Stp/space.h", line 63: Syntax error
"/var/opt/K/SCO/link/1.1.0Eb/etc/conf/pack.d/Stp/space.h", line 65: warning: siz
eof returns 0
"/var/opt/K/SCO/link/1.1.0Eb/etc/conf/pack.d/Stp/space.h", line 70: Syntax error
ERROR: '/var/opt/K/SCO/link/1.1.0Eb/etc/conf/pack.d/Stp/space.c' will not compile properly

i386ld space.o fatal: Can't open file /var/opt/K/SCO/link/1.1.0Eb/etc/conf/pack.
d/Stp/space.o for input
ERROR: Can not link-edit unix

idbuild: idmkunix had errors.
System build failed.

While adding this patch, the system crashed initially. Well halts actually, no errors. Several times later I am able to add the first part of the patch. While it is working on the second part it halts again, it comes back up and I get the following. Trying to run Custom.

The full error:
Fatal error: building package index for `/opt/K/SCO/SoftMgr/1.0.1a/custom/custom.tlib
Error Information logged to: /tmp/tclerror.795.log

What /tmp/tclerror.795.log has:

-------------------
Error code = NONE
--------------------
No "#@package:" definitions found in /opt/K/SCO/SoftMgr/1.0.1a/custom/custom.tli
b
while executing
"error "No \"#@package:\" definitions found in $libName""
invoked from within
"if {$packageCnt == 0} {
error "No \"#@package:\" definitions found in $libName"
} "
invoked from within
"if {[catch {
scanfile $contectHdl $libFH
if {$packageCnt == 0} {
error "No \"#@package:\" definitions found in $libName"
..."
(procedure "TCLSH:CreateLibIndex" line 35)
invoked from within invoked from within
"TCLSH:CreateLibIndex $libfile"
invoked from within
"if {$status != 0} {
global errorInfo errorCode
error "building package index for `$libfile' failed: $errmsg" $erro
rInfo $erro ..."
("foreach" body line 5)
invoked from within
"foreach libfile $libfilelist {
set status [catch {
TCLSH:CreateLibIndex $libfile
} errmsg]
if {$status != 0} {
..."
(procedure "buildpackageindex" line 2)
invoked from within
"buildpackageindex /opt/K/SCO/SoftMgr/1.0.1a/custom/custom.tlib"
while loading Tcl library index "/opt/K/SCO/SoftMgr/1.0.1a/custom/custom.tnd
x"
invoked from within
"loadlibindex $customTlib"
(procedure "PrepEnv" line 40)rInfo $erro ..."
("foreach" body line 5)
invoked from within
"foreach libfile $libfilelist {
set status [catch {
TCLSH:CreateLibIndex $libfile
} errmsg]
if {$status != 0} {
..."
(procedure "buildpackageindex" line 2)
invoked from within
"buildpackageindex /opt/K/SCO/SoftMgr/1.0.1a/custom/custom.tlib"
while loading Tcl library index "/opt/K/SCO/SoftMgr/1.0.1a/custom/custom.tnd
x"
invoked from within
"loadlibindex $customTlib"
(procedure "PrepEnv" line 40)
invoked from within
"PrepEnv"
("uplevel" body line 33)
invoked from within
"uplevel $command"
===============================

Now comes the fun part... The HW tech doing this work has had a history of making things worse when he leaves, last time it was disconnect the CD-ROM. On this occasion (and the reason for rebuilding the kernel initially) he had the SCSI ID of the TD set for 3 when the original was set for 2 and this conflicted with the HD. Now I don't want to put the blame on him for this issue, but weird things happen when he is around. They also experienced a memory error that was not caught prior to the problem of the TD. So I am uncertain whether I am looking for a HW issue or a software issue. But I speculate HW issues which now have damaged SW. I am thinking the CUSTOM.TLIB file is an indexing file of what is currently installed, could someone please confirm and offer any help as to repair possibilities?

Thanks Scott

 

[stanhubble]

Taken from

http://stage.caldera.com/cgi-bin/ssl_reference?105776

This problem was fixed in Release Supplement 504C for SCO OpenServer 5.0.4 (rs504c).
However, if you have rs504c installed on the system and the kernel still fails to relink, follow the instruction below to correct the problem.

In order to correct this problem, run 'mkdev tape' and add a SCSI tape drive. In order to save time, do not choose to relink the kernel. Note that you may be unable to add a second device at the original device parameters. If this is the case, then use a different scsi id, bus or lun to add the drive.

Edit the file /etc/conf/sdevice.d/Stp and delete the second line.

Then, run 'mkdev tape' again and remove the SCSI tape drive. Relink the kernel.