Here is a strange thing. I was upgrading a friend's empeg (I've done several now using the excellent drive upgrade FAQ) and a weird thing happened.

This is what I did:

1. Removed the two old disks.
2. Installed the two new disks.
3. Ran the builder
4. Installed the 2.0 final.
5. Removed one new disk and installed one old disk as secondary.
6. Copied the data.
7. *Tried* to switch drives to copy the other pair.

At this point, the strangest thing happened. The boot logs indicated that the empeg was recognizing the drives, but I got Hard Disk Not Found. Odd. So, I did the following:

1. Triple checked all of the connections.
2. Carefully recrimped all the connectors on the IDE cable.
3. Tried again.

Same schidt. So, I figured I'd try one drive at a time. It would work with neither of the 2nd pair of drives (one old and one new), but it would work just fine with the first new drive I had copied. Very strange. How could *two* drives have gone bad? I was careful with ESD and that whole business. I tried them several times thinking I had some kind of intermittent failure on the cable or connector. Nope -- totally repeatable.

I decided to see if the builder image could find the second new drive (the regular image couldn't, apprarently). So, I did this:

1. Install the second new drive as the only drive.
2. Try it again one last time to make sure it doesn't work.
2. Load the builder image.
3. See what happens.

Well, this is what happened: the builder recognized the disk and even recognized it as already built. I pressed enter to rebuild the disk. With the procedure I just did, I was trying to eliminate physical connectivity issues because I tried it with two images (2.0 final -- didn't work, builder -- did work) without touching or moving the drive or cable at all.

After the disk was rebuilt, I reloaded the 2.0 final. It worked, so I installed the other disk (the remaining old disk) and copied the data.

Its now working with both new disks.

Sorry if this wasn't worth reporting, but it seemed odd and I thought I'd mention it.

Jim

I believe you might be describing a bug that Paul Grzelak (at least I think it was Paul) reported recently with the builder image. Anyone have a link to that thread?

Either that or it's a bad IDE header that's giving intermittent trouble.

2. Installed the two new disks.
3. Ran the builder

Doesn't the builder only build the disk on the primary interface? If so the slave disk wouldn't be built, plus you would also need to install 2.0final on both disks seperatly.

Doesn't the builder only build the disk on the primary interface?

Good question. I thought it built both of them, and I think the D.U.G. even says so. If that's wrong, I need to correct that. Anyone have definitive information?

I must admit I thought I had read that in the D.U.G., but I just went back and checked and it does say both, and as I have never tried it with two disks attached, I am probably wrong.

The problem of the drives not booting is probably due to:
Two old drives - Primary contains files needed to fully boot
Two new drives - 2.0 only put onto Primary
Boot with New master & old master set to slave - both drives contain the files needed to boot, so everything works.
Boot with new slave set to master and old slave - neither disk has had a SW version applied, so boot fails due to missing files.

Just a guess

I just went back and checked and it does say both, and as I have never tried it with two disks attached, I am probably wrong.

I could just as easily be the one who's wrong. I also have never tried to use the builder to format two disks at the same time. Anyone know for sure?

Boot with new slave set to master and old slave - neither disk has had a SW version applied, so boot fails due to missing files.

Hey, here's a good question. Anyone who knows the kernel software and is willing to look at the source code could answer this...

If you have a properly partitioned drive, but no player software installed on that drive, does the kernel display the "no hard disk found contact support" message, or does it display some other message?

Greetings!

Is this the thread you were looking for?

Hey I had the same thing happen to me but it was a while ago and with v2b13 on win98 it was fine after I ran the builder the second time. I did post somthing about it but I cant find it now.

found the post here

the builder builds both drives, if attached. it only stress tests the primary, however.

the drives in a couple of my units were built this way, and are working just fine. the upgrade guide is accurate as is, i believe.

--dan.

Cool, thanks! Always nice to see news that means I don't have to edit anything.

When watching the builder log, I noticed that it built both drives.

I was under the impression that the player software was loaded onto the flash. I'd really like to understand that better. Is there anything critical on the drives that is required at boot time *and* is not installed by the builder on to the secondary drive?

One thought: the builder might build primary and secondary differently. If so, when replacing both drives and copying from two old drives, one would need to run the builder on the new drives twice, once with each drive as the primary.

The kernel is stored in flash. The player software (and other OS stuff) is stored on the hard drive. When you install new player software, it installs both, but only installs the player software on the primary drive. If, for some reason, the drives were swapped, the kernel would definitely remain the same, but the player software would be whatever was on the old secondary drive. This could be the same, an older version, or nothing. Either of the last two is pretty bad. Of course, the drives aren't going to magically swap on their own.

Also, after you've done the builder thing, the drives have no software on them. They've really just been partitioned in preparation for software (and music).