The BIOS for my server doesn't have the option of "remembering" the disks on the IDE channel. The only option is "Auto", which causes it to probe the disks each time.

Having swapped a few disks around, I've noticed that certain combinations of disks causes the probe to stall, after identifying the primary disk, for around 30 seconds, before finding the secondary.

It's always when the primary drive is an "older" disk than the secondary. Switching the disks around on the channel cures it. I've tried with a number of disks, and also cables, so it's not that causing it. Both drives are jumpered for Cable Select, and the cables are all 80-wire.

Currently, I have a 20Gb IBM Deskstar, ATAPI-4, as the primary. And a 250Gb Maxtor, ATAPI-7, as the secondary. Previously, I had the Maxtor as the primary, and a 500Gb Seagate as the secondary, and this also exhibited the same symptoms.

Does anyone have any suggestions, other than switching the drives on the channel, that might stop this "delay".

Cheers.

Check the label on the top of each disk to be sure about the jumper setting. I have seen some disks that do not have the usual settings. For example I have a Maxtor IDE disk that has to have "cable select" if the disk is the only one on the controller. On that disk you only use the master/slave jumpers if there are two disks on the controller.

On a different note I've had very bad experience with IBM Deskstar disks. I purchased six at one time. Four died within a year.

I double checked the jumper settings on the label, and also on the web site.

I was going to try master/slave, but I've not had to do that for so long. Everything has always worked fine with CS.

The Deskstar has been the OS drive in one of my servers for ever it seems. Guess if it's lasted this long, it must have been one of the "golden" ones.

Cheers.

Well, switching the drives to Master/Slave didn't make any difference. It still "pauses" during the autodetect.

Cheers.