OpenServer 5.0.5 Compaq software shutting down server 1

[nashcom]

I’ve got OpenServer 5.0.5 running on an old Compaq Proliant 400 server. It was installed around 1999.

There seems to be a problem with the main-board or riser-board related to the system Fan 1 (on the backplane of the server behind the CPU). It’s not getting up to speed. I’ve tried another fan from a ProSignia server (same fan model) and that’s doing the same. When I connect either fan to the Fan2 socket on the riser-board the fan spins at the correct speed, but the system needs to have Fan1 running for it to work.

The system boots from the Compaq SmartStart CD or from a Windows98 floppy and remains up.

However, when booting into OpenServer, cpqw (Compaq Wellness) driver reports that Fan1 has died, and says it will shut down the system after one minute….which it does!

It does this if it’s booted into Single-User mode too.

From what I can see, it’s loading something called cpqwinit at the time it shows the ‘hwconfig’ screen, and before it loads the SCSI HBA drivers etc. I suspect that ‘cpqwinit’ is loading cpqw.

What I’d like to do is disable the Compaq monitoring software from loading at SCO startup. I know it’s not a good idea to run the server without a fan, but it’s a Pentium II processor with a huge heatsink (and no CPU fan); the room it’s in doesn’t get hot; and I thought I could temporarily run the server with the side panel removed and a desk fan blowing air into it.

However, I don’t know how to disable the Compaq stuff from startup. I’m not sure if it’s built into the kernel, or if it somehow loads from the /etc/rc?? scripts. As I say, I only have less than one minute from the time of booting into single-user mode and the server starting to shutdown, so I really need to know what I’m doing. I’m going to try to get to another site that’s running OpenServer 5.0.5 on a Proliant 200 server, and see if I can work out how to disable the Compaq stuff, although I’m not sure if the same drivers are on that server.

I did see that there’s a /dev/cpqw file, but the attributes were listed as something like crwx – I’m not sure what the ‘c’ signifies, and I wasn’t able to copy the ‘file’. I’d wondered if I could copy it to a different name and then remove it.

Any help would be most appreciated!

Thanks

David

 

[Annihilannic]

The 'c' signifies a 'character' device.

The safest bet would be to move it to /dev/cpqw.disabled, for e.g., that way you could easily restore it if necessary.

Annihilannic.

 

[nashcom]

Thanks Annihilannic.

When I tried to copy the file it said Error reading data from cpqw no such device (error 19). How should I move it?

I've been reading a bit more about it. It seems that the cpqw Wellness driver is linked into the Kernel, and the recommended way to remove it is 'mkdev cpqhealth' then relink the kernel. I don't have that option as the server isn't staying up long enough.

I'm thinking that I could do the following:

At the Boot: prompt, tell the system to boot from a backup kernel in /stand. I'm not sure what to type for the boostring but hopefully I should be able to work it out.

In /etc/conf/sdevice.d there's a file called cqpw that looks like

cpqw Y 1 0 0 0 0 0 0 0

I thought I could edit this file and change the 'Y' to a 'N'.

If I then re-link the kernel (etc/conf/cf.d/link_unix -y) will this give me the exact kernel I've been running with the exception that cpqw is removed?

I'm going to sleep on it and try it in the morning.

Again, it'd be great if someone could read thorough what I'm proposing to do, and let me know if it sounds okay.

Thanks

David

 

[PHV]

How should I move it?
mv /dev/cpqw /dev/cpqw.disabled

Hope This Helps, PH.
Want to get great answers to your Tek-Tips questions? Have a look at FAQ219-2884 or FAQ181-2886

 

[nashcom]

Thanks again!

I actually managed to do it a bit more simply. I entered the following at the boot prompt:

defbootstr disable=cpqw

This then let me boot the server without the Wellness driver.

I've decided not to permanently remove the device from the kernel, at least, not for now. We'll try to source replacement hardware.

In the meantime I've connected the fan to the Fan2 port which means the system is being cooled properly. The only problem is that we won't be alerted to future failure (because the cpqw driver isn't being loaded).

Thanks for your help

David