just wanted to share my experiences. i finally got around to installing fc5 today on a fairly modern pc. during the installation i encountered a serious problem which i had not had while installing ubuntu 5.10 or suse 10.0 on the same computer. now that i've solved the problems, i thought i'd tell you, in case anybody else is in the same situation. basically the computer had 2 IDE hard drives and 2 cd/dvd drives. they were arranged so that one hard drive and one cd or dvd were connected to each IDE port on the motherboard. there was a seemingly random arrangement of slave/master/cable-select jumpers which appear to be the reason for the failing installation. now i've reduced the system to one hard-drive (master) and one dvd-drive (slave) both connected to the primary IDE-port, and the installation is running fine. before opening the computer and changing the internal structure, i did of course try to rectify the problem using the bios, but this didn't help much. so if you have problems with an FC5-installation, check your jumper settings hold

adjust you title to show:

fc5 install troubles: fixed

or better - state what it was that anaconda said made it fail.

the body of the post needs to include the symptoms of the failure - what was it that went wrong and when? How would someone else recognise this? (You know - the sort of details someone would search for?)

AND: did you try explicitly setting the ide device jumpers to master/slave (or setting them all to cable select) as required?

wish i could. the messages flew past too quickly for my weak eyes to read after selecting install from the boot screen. with carefull observation however, it became clear that the attempt to acknowledge the hard-drives was producing a long error message. as said, this dissapeared pretty quickly.


symptoms: installation froze at random moments. sometimes it managed to almost install the first cd, at other moments it froze before the x-windows server for the graphical install could start. text-install also froze.

i should have written down what the jumper settings were beforehand, because i've forgotten now. what they are now is:

i hard-disk and 1 dvd-rom, both on IDE-primary, hard-disk as master, dvd-rom as slave.

curious is the fact that suse 10.0 and ubuntu 5.10 (or probably better said the options used while compiling the kernel used while booting in these two cases) coped without batting an eyelid.

i hope this at least clarifies how little i know about the problem

hold

Yeah - that pretty much does the trick.

having accused the jumper settings of being the root of all evil, i may have to eat humble pie and at the same time direct my accusations at another culprit.

the thing is, the first thing i did when i got the computer was take it apart to see what was inside. then i didn't put the case back on but tried to install fc5 with an open case. that didn't really bother me at the time, but i was at a friend's house recently trying to do the same thing and the install didn't work there either.

at some stage, i brushed against the cd-drive and almost burnt myself. these babies get hot. apparently the circulation in a closed case is beneficial, so i'll try putting the lid back on my fc box and installing again. if it works, i'll know why it was going wrong.

hold

Yeah, I've heard that too - not sure why: it is not exactly airstreamed in there... I suspect it messes with the forced convection part. However, if the room has good airflow/air-conditioning, it should be better with the case open.

Of course, there's aways sticking a column of liquid Helium on the mobo...

it suddenly occured to me, that the reason is really simple. the computer has to know how warm it is. in one computer i saw there was actually a small cabel hangin down from the motherboard to measure the temperature. if there isn't any circulation, the computer can't find out how hot it is and doesn't power up the fan. that's why the closed casing is a good idea

My DT system cases (server or fullATX) are just the skeleton, left Wide open for HW modifications. I swap out & test various RAID (Siig, Via) controllers on; PCI add-in cards, PCI Express x1 cards. Normally NO heat problems on AMD 939 nForce4 mainboards (Abit, GA, SL, or Tyan). Had to provide extra cooling for DDR400 RAM (Crucial) while testing stable overclocking. I OC'd my other DDR400 GB modules to ~504 stable. LT always COOL!
I also sometimes direct Xtra fan to GPU during OC tests, as most current GPU (ATX design) cards are poorly designed for convection cooling.
S B! Any idea exactly What HW h_b has? I have some newer HD (16MB buffer or SATA 3.0) that get hot=overheat anything that is above them. FIX=Mount external and phasechange cool!