Tech Support Journal

System startup with ‘Pci.sys is missing or corrupt’.

An attempt to boot from XP CD generates an error ‘File setupdd.sys could not be loaded.
The error code is 14
Setup cannot continue.’

Removing the pci modem changes the error code from 14 to 7. Removing and reinstalling the AGP graphic card and memory modules and ensuring that they are fitted properly boots the system up fine without the need for a reinstallation of Windows.

There are also other things to bear in mind when you’re planning to troubleshoot or recover and XP crash, (These points are especially useful when your system crash with the error - “Windows could not start because the following file is missing or corrupt C:\windows\system32\config\system”) ;

1. There is always a folder in Windows installation folder that consists system files that can restore your system to the state of your first install, provided that the crash isn’t hardware failure related. (e.g. C:\windows\repair) The folder with your system files prior to the crash is located at c:\windows\system32\config. Copying files (after backing up the original files) to the above location will restore your system to first boot state.

2. Enabling wildcards to ease copy and deletion process, which can be useful when you’re working in recovery console mode. The command is  AllowWildCards = TRUE , to be typed at the command prompt. Another command AllowAllPaths = TRUE is also useful as it allows you to access all files and folders on the system.

3. Once you’re back into your first install state, you can then set explorer to display hidden and system files and folders, disable simple file sharing from the folder options in Windows explorer. This will enable you to see a folder call “System Volume Information Folder” which consist system files backup at various time from c:\windows\system32\config . You may need to give yourself rights to the folder in order to access its contents.

Restoring the system files from point 3 will be a closer match to your last system state prior to the crash.

A system that was working fine all along suddenly starts rebooting itself & was unable to boot into Windows XP GUI interface.

Starting the system in debug mode without restart, an option that you can select by pressing F8 key before the startup process, shows an unmountable boot volume error.

Booting with an XP boot CD & doing a directory list with dir command shows an error like unable to enumerate drive.

Reinstalling Windows XP on another drive so as to retain the data was the first option taken but the problem was finally resolved by just replacing the power supply unit which seems  to have a faulty fan causing the POST test to proceed slower than usual.

System boot up without any error immediately after replacing the PSU.

A simple task of replacing a burnt 3COM switch & rearranging the patch cords in proper sequence turns into a major troubleshooting process when after plugging in the new switch & cascading it to an existing switch, all users on the network cannot no longer access the network.

It is no help that these two switches are also linked via a fiber link to another department nearby.  The fact that the face plates at the users’ locations are just partially numbered & the patch cords aren’t numbered at all adds on to the difficult task of isolating the problem.

First suspect was that there was a loop somewhere whereby there are two uplinks linking up two of the switches, namely the fibre link & another ethernet link which we have accidentally plugged in to the patch panel while disconnecting & rearranging the cables. But because there were no indication nor any kind of network diagram, we have to test it one by one;

by first disconnecting the all the cables on the server rack on our end,

cascade the two switches, then connect the fiber link,

ensure the remote location is able to access the network,

then connect a workstation to the local switches doing at command prompt, an ipconfig /release then ipconfig /renew every time a patch cord is connected from the patch panel on the server rack to the port on the switch.

This is to test whether to ensure we can obtain a IP lease from the server without any problem. If we can’t then that is the source of the problem.

The process also force us to properly label the patch cords, tidy up the cable layout & finally isolating the problematic point on the patch panel, all these while with the users breathing down our necks.

Once all these are done & the network is fully functional again,  there is still one particular user that is unable to access the network. In fact the network icon on her computer shows a network disconnected sign.

Suspect that it could be a non-functioning port on the old switch that we have cascaded the new switch to since we have done a rearrangement of the network points.

Without a data point number on the face plate near her table, we have to unplug those ports that is not in use from the patch panel (that is their led is not lit on the old switch) then plug one end of a network cable to a port on the new switch & then plug the other end to the patch panel unused ports one by one.

Since those ports are unused, there should not be any led light on the new switch when we plug the cable to each of the patch panel’s port. If the led is lit, we have isolated the faulty port on the old switch.

We manage to isolate the problematic port on the old switch, tape it up to ensure no one would accidentally use it again & check every single network point to ensure it is functionally.

The whole process took 2.1/2 hours but it is better than having to deal with errantic problems these two issues may cause which we may be clueless about few weeks down the road.