SEO

Greetings all !

I have just cloned a Toshiba Tecra laptop
running SCO Open Server 4.0.5 (and Win 95)
using the DD command in a Linux forensics distribution.
NB ! we have SCO licenses for two (and more)
machines but chose to clone since the SCO/Win95
coniguration & tools are very special and complex to
install from scratch.

The two laptops were the exact same
model, same harddisks, network card etc
and the cloning was successful.
SCO runs fine and so does Win95.

This laptop is special as it is used to
run different tasks on ticket vending machines.

We then tested both laptops on a routine
job in a ticket vending machine and the original
machine did the task as it should (installation
of the ticketing machine), but the
cloned machine failed.

I have had a look through both machines
and have found the following differences :

The cloned machine has a defect CD rom
and the battery lamp is yellow.
This is also reflected during boot
as the CD rom is not mounted on the cloned machine.
(CD ROM is not used for the task against
the ticketing machine)

I also checked Process Manager and the
original machine has 46 procs while
the cloned machine has 42.

The missing processes on the cloned machine were :

initd
routed
snmpd
strerr

Also the PID's are different on the two
machines but this should be no problem.

During the task we tested with the laptop
connected to the ticketing machine.
The ticketing machine is booted via a floppy
while SCO is running,
the ticketing machine SW then should
fetch data from the SCO laptop but this seemed to fail.
So a missing service might be the cause.

Thw network lamp is green and SCO does
not give any network error during boot,
I did also try to swap network cards.

If anyone have any comments on the missing Procs
on the cloned machine or any tips on where
to check & compare the two laptops this would
be greatly appreciated.

Thanks alot !

kind regards,

M~

Another Visitor wrote:
> Greetings all !
>
> I have just cloned a Toshiba Tecra laptop
> running SCO Open Server 4.0.5 (and Win 95)
> using the DD command in a Linux forensics distribution.
> NB ! we have SCO licenses for two (and more)
> machines but chose to clone since the SCO/Win95
> coniguration & tools are very special and complex to
> install from scratch.
>
> The two laptops were the exact same
> model, same harddisks, network card etc
> and the cloning was successful.
> SCO runs fine and so does Win95.
>
> This laptop is special as it is used to
> run different tasks on ticket vending machines.
>
> We then tested both laptops on a routine
> job in a ticket vending machine and the original
> machine did the task as it should (installation
> of the ticketing machine), but the
> cloned machine failed.
>
> I have had a look through both machines
> and have found the following differences :
>
> The cloned machine has a defect CD rom
> and the battery lamp is yellow.
> This is also reflected during boot
> as the CD rom is not mounted on the cloned machine.
> (CD ROM is not used for the task against
> the ticketing machine)
>
> I also checked Process Manager and the
> original machine has 46 procs while
> the cloned machine has 42.
>
> The missing processes on the cloned machine were :
>
> initd

do you mean inetd ?

> routed
> snmpd
> strerr
>
> Also the PID's are different on the two
> machines but this should be no problem.
>
> During the task we tested with the laptop
> connected to the ticketing machine.
> The ticketing machine is booted via a floppy
> while SCO is running,
> the ticketing machine SW then should
> fetch data from the SCO laptop but this seemed to fail.
> So a missing service might be the cause.
>
> Thw network lamp is green and SCO does
> not give any network error during boot,
> I did also try to swap network cards.
>
> If anyone have any comments on the missing Procs
> on the cloned machine or any tips on where
> to check & compare the two laptops this would
> be greatly appreciated.

perhaps the network card is not getting the exact same pci bus address & irq
as the on the old machine even though everything is nominally the same. this
is common.

run netconfig on the new box, highlight the nic, hardware, modify, update
those values
get the correct values from "hw |pg" (also on the new box)

Here's a little script to extract just what you are looking for out of all
the other output from the hw command

Run this on both the old and new machines:

hw -r pci |awk '/Index:/{E=0};/DeviceNum:/{N=$0};/Function:/{F=$0}
/Bus:/{B=$0};/VendorId:/{V=$0};/DeviceId:/{M=$0};/Ethernet/{E=1}
/InterruptLine:/{I=$0;if(E==1) print V"\n"M"\n"I"\n"B"\n"N"\n"F"\n";E=0}'

It should generate a paragraph like this for each ethernet card installed:

VendorId: 3Com Corporation
DeviceId: 0x9200
InterruptLine: IRQ-5
Bus: 3
DeviceNum: 2
Function: 0


are the results different on the new machine from the old machine?

Assuming they are different, write down the above values from the new
machine
Then run "netconfig" on the new machine, highlight the nic, hardware,
modify, and change the values to match the results you captured from above.
answer yes to the questions about relinking the kernel and rebuilding kernel
environment and reboot.

--
Brian K. White -- brian@aljex.com -- http://www.aljex.com/bkw/
+++++[>+++[>+++++>+++++++<<-]<-]>>+.>.+++++.+++++++.-.[>+<---]>++.
filePro BBx Linux SCO Prosper/FACTS AutoCAD #callahans Satriani

"Brian K. White" <brian@aljex.com> wrote in
news:CJCdnQuHH-x-SfbdRVn-jg@comcast.com:

> Another Visitor wrote:
>> Greetings all !
>>
>> I have just cloned a Toshiba Tecra laptop
>> running SCO Open Server 4.0.5 (and Win 95)
>> using the DD command in a Linux forensics distribution.
>> NB ! we have SCO licenses for two (and more)
>> machines but chose to clone since the SCO/Win95
>> coniguration & tools are very special and complex to
>> install from scratch.
>>
>> The two laptops were the exact same
>> model, same harddisks, network card etc
>> and the cloning was successful.
>> SCO runs fine and so does Win95.
>>
>> This laptop is special as it is used to
>> run different tasks on ticket vending machines.
>>
>> We then tested both laptops on a routine
>> job in a ticket vending machine and the original
>> machine did the task as it should (installation
>> of the ticketing machine), but the
>> cloned machine failed.
>>
>> I have had a look through both machines
>> and have found the following differences :
>>
>> The cloned machine has a defect CD rom
>> and the battery lamp is yellow.
>> This is also reflected during boot
>> as the CD rom is not mounted on the cloned machine.
>> (CD ROM is not used for the task against
>> the ticketing machine)
>>
>> I also checked Process Manager and the
>> original machine has 46 procs while
>> the cloned machine has 42.
>>
>> The missing processes on the cloned machine were :
>>
>> initd
>
> do you mean inetd ?
>
>> routed
>> snmpd
>> strerr
>>
>> Also the PID's are different on the two
>> machines but this should be no problem.
>>
>> During the task we tested with the laptop
>> connected to the ticketing machine.
>> The ticketing machine is booted via a floppy
>> while SCO is running,
>> the ticketing machine SW then should
>> fetch data from the SCO laptop but this seemed to fail.
>> So a missing service might be the cause.
>>
>> Thw network lamp is green and SCO does
>> not give any network error during boot,
>> I did also try to swap network cards.
>>
>> If anyone have any comments on the missing Procs
>> on the cloned machine or any tips on where
>> to check & compare the two laptops this would
>> be greatly appreciated.
>
> perhaps the network card is not getting the exact same pci bus address
> & irq as the on the old machine even though everything is nominally
> the same. this is common.
>
> run netconfig on the new box, highlight the nic, hardware, modify,
> update those values
> get the correct values from "hw |pg" (also on the new box)
>
> Here's a little script to extract just what you are looking for out of
> all the other output from the hw command
>
> Run this on both the old and new machines:
>
> hw -r pci |awk '/Index:/{E=0};/DeviceNum:/{N=$0};/Function:/{F=$0}
> /Bus:/{B=$0};/VendorId:/{V=$0};/DeviceId:/{M=$0};/Ethernet/{E=1}
> /InterruptLine:/{I=$0;if(E==1) print
> V"\n"M"\n"I"\n"B"\n"N"\n"F"\n";E=0}'
>
> It should generate a paragraph like this for each ethernet card
> installed:
>
> VendorId: 3Com Corporation
> DeviceId: 0x9200
> InterruptLine: IRQ-5
> Bus: 3
> DeviceNum: 2
> Function: 0
>
>
> are the results different on the new machine from the old machine?
>
> Assuming they are different, write down the above values from the new
> machine
> Then run "netconfig" on the new machine, highlight the nic, hardware,
> modify, and change the values to match the results you captured from
> above. answer yes to the questions about relinking the kernel and
> rebuilding kernel environment and reboot.
>



Thanks alot for your follow-up !
I had a chance to test this just recently and
the results were as follows.

netconfig & hw |pg confirmed that the hw & network values
are similir on both machines.

The script did not seem to work for me on this laptop,
I ran it in a Unix window from SCO and it either just stood
working for a while and then a new #prompt came up.
Or it went into > mode.

But I tried to swap the harddisks just to see if
we might have any HW issues and this did not solve the problem,
so it might not be HW related after all.

I am npw trying to figure out exactly what the boot floppy in
the ticketing machine is supposed to fetch from the SCO laptop,
I am also curious about the processes (Procs) that are not running
on the cloned machine.

These were :

routed
snmpd
strerr
inetd (sorry, not initd indeed)

Might these be related and how can I start them manually ?

thanks again & cheers,

M~

Another Visitor wrote:
> "Brian K. White" <brian@aljex.com> wrote in
> news:CJCdnQuHH-x-SfbdRVn-jg@comcast.com:
>
>> Another Visitor wrote:
>>> Greetings all !
>>>
>>> I have just cloned a Toshiba Tecra laptop
>>> running SCO Open Server 4.0.5 (and Win 95)
>>> using the DD command in a Linux forensics distribution.
>>> NB ! we have SCO licenses for two (and more)
>>> machines but chose to clone since the SCO/Win95
>>> coniguration & tools are very special and complex to
>>> install from scratch.
>>>
>>> The two laptops were the exact same
>>> model, same harddisks, network card etc
>>> and the cloning was successful.
>>> SCO runs fine and so does Win95.
>>>
>>> This laptop is special as it is used to
>>> run different tasks on ticket vending machines.
>>>
>>> We then tested both laptops on a routine
>>> job in a ticket vending machine and the original
>>> machine did the task as it should (installation
>>> of the ticketing machine), but the
>>> cloned machine failed.
>>>
>>> I have had a look through both machines
>>> and have found the following differences :
>>>
>>> The cloned machine has a defect CD rom
>>> and the battery lamp is yellow.
>>> This is also reflected during boot
>>> as the CD rom is not mounted on the cloned machine.
>>> (CD ROM is not used for the task against
>>> the ticketing machine)
>>>
>>> I also checked Process Manager and the
>>> original machine has 46 procs while
>>> the cloned machine has 42.
>>>
>>> The missing processes on the cloned machine were :
>>>
>>> initd
>>
>> do you mean inetd ?
>>
>>> routed
>>> snmpd
>>> strerr
>>>
>>> Also the PID's are different on the two
>>> machines but this should be no problem.
>>>
>>> During the task we tested with the laptop
>>> connected to the ticketing machine.
>>> The ticketing machine is booted via a floppy
>>> while SCO is running,
>>> the ticketing machine SW then should
>>> fetch data from the SCO laptop but this seemed to fail.
>>> So a missing service might be the cause.
>>>
>>> Thw network lamp is green and SCO does
>>> not give any network error during boot,
>>> I did also try to swap network cards.
>>>
>>> If anyone have any comments on the missing Procs
>>> on the cloned machine or any tips on where
>>> to check & compare the two laptops this would
>>> be greatly appreciated.
>>
>> perhaps the network card is not getting the exact same pci bus
>> address & irq as the on the old machine even though everything is
>> nominally the same. this is common.
>>
>> run netconfig on the new box, highlight the nic, hardware, modify,
>> update those values
>> get the correct values from "hw |pg" (also on the new box)
>>
>> Here's a little script to extract just what you are looking for out
>> of all the other output from the hw command
>>
>> Run this on both the old and new machines:
>>
>> hw -r pci |awk '/Index:/{E=0};/DeviceNum:/{N=$0};/Function:/{F=$0}
>> /Bus:/{B=$0};/VendorId:/{V=$0};/DeviceId:/{M=$0};/Ethernet/{E=1}
>> /InterruptLine:/{I=$0;if(E==1) print
>> V"\n"M"\n"I"\n"B"\n"N"\n"F"\n";E=0}'
>>
>> It should generate a paragraph like this for each ethernet card
>> installed:
>>
>> VendorId: 3Com Corporation
>> DeviceId: 0x9200
>> InterruptLine: IRQ-5
>> Bus: 3
>> DeviceNum: 2
>> Function: 0
>>
>>
>> are the results different on the new machine from the old machine?
>>
>> Assuming they are different, write down the above values from the new
>> machine
>> Then run "netconfig" on the new machine, highlight the nic, hardware,
>> modify, and change the values to match the results you captured from
>> above. answer yes to the questions about relinking the kernel and
>> rebuilding kernel environment and reboot.
>>
>
>
>
> Thanks alot for your follow-up !
> I had a chance to test this just recently and
> the results were as follows.
>
> netconfig & hw |pg confirmed that the hw & network values
> are similir on both machines.
>
> The script did not seem to work for me on this laptop,
> I ran it in a Unix window from SCO and it either just stood
> working for a while and then a new #prompt came up.
> Or it went into > mode.

the time it went to ">" you didn't type it right. (missed a quote or brace
or something)
the time it went to # you typed it wrong but different than before, or the
card is not a pci card (or on the pci bus via pcmcia) or the output from
5.0.4 hw command is different than 5.0.5 and 5.0.6 which are the boxes where
I wrote it and tested it, or your card is not being detected by the kernel
at the hardware level for some reason, or it just plain failed to recognize
the stanza for your network card for whatever reason. In any event it was
just something I wrote off the top of my head for that post and not really
important. It was just an attempt to describe what to look for in "hw" since
95% of hw's output does not matter in this case.

> But I tried to swap the harddisks just to see if
> we might have any HW issues and this did not solve the problem,
> so it might not be HW related after all.
>
> I am npw trying to figure out exactly what the boot floppy in
> the ticketing machine is supposed to fetch from the SCO laptop,
> I am also curious about the processes (Procs) that are not running
> on the cloned machine.
>
> These were :
>
> routed
> snmpd
> strerr
> inetd (sorry, not initd indeed)
>
> Might these be related and how can I start them manually ?

All of those services are started by /etc/tcp
/etc/tcp is run at start up b virtue of the symlink to it named
/etc/rc2.d/S85tcp
(does /etc/rc2.d/S85tcp exist?, are you starting in runlevel 2?)

what happens on the new box when you manually run this? (as root of course):
/etc/tcp start

on both boxes, run these commands and compare the results:
ifconfig -a
netstat -rn
netstat -i
netstat -m

try those both before and after manually running "/etc/tcp start" and look
for differences there too

and you already compared "hw |pg" and it showed the card being detected on
the exact same irq, bus, device & function? That is actually not too common.

the fact that netconfig looks the same doesn't mean much. netconfig is not
reading the card, it just reading the kernel config which only changes when
you change it.

netconfig sometimes autodetects the card if you are installing a supported
card for the first time and there happens to be no irq conflict. You could
test that by using netconfig to uninstall the card, relink & install kernel
& reboot, then netconfig again and "add new lan adapter" and see if it
detects it again, but that will lose you your network settings (ip address,
netmask, etc...) so you'd have to know everything that matters about the
networking set up so you can set it back up from scratch. So that's why I
haven't said to do that yet, although that is probably what _I_ would do if
I were at the machine myself and had looked over the settings first and
satisfied myself that I knew what was important. I wouldn't want you to
blast your settings and me not be able to help you replace them since it
could take a long time via newsgroup.

Just an aside:
routed you almost certainly don't want on either box. It's there just
because the OS ships with it enabled by default, but on all my boxes I
disable it (on boxes older than 5.0.7 by renaming /etc/routed to
/etc/routed.disabled. Basically routed listens for rip packets and if any
are heard, modifies the kernel routing table accordingly. However usually
your start scripts set up the correct routing table and the last thing you
want is for anything to change it.

good luck
--
Brian K. White -- brian@aljex.com -- http://www.aljex.com/bkw/
+++++[>+++[>+++++>+++++++<<-]<-]>>+.>.+++++.+++++++.-.[>+<---]>++.
filePro BBx Linux SCO Prosper/FACTS AutoCAD #callahans Satriani