bandwidth strange fn of HD
 

here's one I can't figure out:

First, I'm running RH 7.2 with pure-ftp serving 2 drives, one an IBM 60GB and another Maxtor 60GB (both about a year old). Storing data on the IBM drive (transferring from (any other[NT4, win2k, linux]) box to linux ftp server) works fine at ~6000kb/s. however storing data on the Maxtor drive starts ~6000kb/s and linearly slows down to about 500kb/s over ~3 minutes. I know the Maxtor drive works better than this since files copied from the IBM drive to the Maxtor drive at the linux prompt transfer at ~6000kb/s. Swapping the cables between the two drives doesn't change anything. Downloading from both drives is flawless. Also, this network is on a hub and the collision light is pretty active when u/l to both drives. When the u/l slows down, the whole linux machine is unresponsive over the network from the box initiating the u/l. the IBM drive is ext3 and the maxtor is ext2. Any ideas on why the maxtor drive (only) exhibits this behavior? I'm wondering if something got corrupted somewhere so reformatting all the drives and starting over would fix it but that's pretty time intenive. if not that then prob just toss the Maxtor...???

thx if any suggestions, - bp

Both drives should be S.M.A.R.T. capable so you should see if you have 'smartd' or 'smartctl' on your setup. Try 'smartctl -v /dev/hdX' (sub X for whichever device you want to test). At least this will let you see if there is a hardware problem. Could be temperature related (happened to me once), or that the drive is just plain ouched.

I trust that there are no other devices on the IDE chain that these drives are connected to.

I can't see the ext3/ext2 formats causing that much disparity.

thx!
 

you were right, it was a heat issue. this is the first time i've had this problem and thought my athlon would need better cooling way before my old P1-200 did. thx for the suggestions mcleodnine, and for saving 60GB of server space.