Houston, we had a problem

After the upgrades today, I was quite disappointed with disk speed. It seemed no matter what I did, the generic-ide module was taking over my drives before the specific (piix) module could load them. While it did allow me to access my data, it was extremely slow, and I couldn’t turn on DMA, no matter what. After much headbanging, several reboots testing different options and attempting every known way to load the module, I finally went back to my last kernel which worked perfectly.

It appears the 2.6.15-1 and associated udev scripts have a problem where ide-generic is loading first, no matter what. I found some debian threads where they want to get this fixed (they are in the -5 range) by .16. Hope they do.

Back Up

System is back up and running with a new drive. (Hey, did you know you can fit 5 3.5″ drives in a 3u that wasn’t designed for it?) 🙂 Definately can’t handle any more in terms of power needs. Next time I need to grow, It’ll be time for a new case and everything.

Debian came back perfectly, only issues I had (besides *very* cramped spaces) was my idiocy of always forgetting which end is master and slave on an IDE cable (now that I have to put two disks on one ide channel).

I definitely remember why I didn’t want to be an ‘IT’ specialist 🙂

Downtime redux

So, I’m about to have some planned downtime. Hopefully it will be kept to a minimum. If things work right, I’ll be wedging in another drive in this box and be up again in a few minutes.

Problem is, it’s a rack box, and pretty tight. There is a chance that I might have to take one of my old scsi drives out, breaking the raid it’s in 🙁 Hopefully we can avoid that.

Worst case, things don’t come back up right… bleh..