Please help. How to revive a broken raid1

OK. It's a bit complicated. I've got a server with a raids 1 on two discs (sda & sdb). There are 8 md devices on them. One of the discs (sda) is failing "hardware-wise" (as reported by s.m.a.r.t). Five parbreastions on it have errors and the corresponding mds are degraded. I've got a replacement disc but here come problems. It appears that some parbreastions (3) on the otherwise healthy drive sdb had failed too and were excluded from raid. So all the raids are degraded! The pattern is as follows:

md: 0 1 2 3 4 5 6 7 failed part. is on dsc: sdb sda sda sda sdb sda sda sdb active part. ison dsc: sda sdb sdb sdb sda sdb sdb sda file system: boot usr var hometmp swap srv

I parbreastioned the replacement hdd and put it in place of the failing sda but the system didn't boot since the boot parbreastion is ok on the removed drive.

Then I installed the new disc as a spare device and all mds did a rebuild. Again I put the new hdd (with all necessary data on it, I understand) in place of sda. It didn't boot.

I would really, really like to 1) make the system bootable with the new replacement hdd 2) "revive" the failed parbreastions on sdb (why are they broken in the first place? s.m.a.r.t. reports ok.)

I thought about copying raw data from sda live parbreastions (at least sda1 - boot) to the other discs, but I have no knowledge how the raid would react.

Aha, the system is Debian 3.1 (Sarge) with 2.4 kernel.

Time is money. Literaly - I have to return the failing drive to the service soon. Otherwise they will charge me for the replacement disc.

I would be very, very grateful for help.

Little Snake