most people here have experience with stuff that's within a typical (power user/enthusiast) user's budget, whereas the folks in the serverfault chat have experience with many-thousand-dollar (new) servers, usually dual or quad processor (or more), hundreds of gigs of ram, etc
I've used hardware RAID cards though, so I could help you either way probably
So here it goes. I have the original RAID card in the machine, and it normally boots fine to an "insert disk" message, when I have no disks in the machine. The second I pop a SATA or SAS disk into any of the bays, I immediately get a pretty long hang from the controller kernel.
I found some light oxidation on pins, cleaned it, and everything seemed to be fine with that hardware.
have you tried the drives individually on the mobo's SATA controller with the same cables and power hookups to be sure that it can at least recognize them in the BIOS? (obviously won't be able to boot due to lack of RAID)
I have yet to tear out more hardware so I have physical access to any SATA ports on the mobo. As far as I've seen there are none accessible without tearing out the IO card(which seems to physically block, and also has the VGA port I need)
"I plugged in each drive individually into the SATA/SAS controller and was able to see it in the BIOS and the BIOS didn't freeze like the controller kernel" would be a good thing for you to be able to claim before I'm convinced that it's the RAID controller or the PCIe bus
I have some sort of card that appears to be the disk controller, with two substantial cables running to the IO card, one for the top set of bays, and another for the bottom.
Different combinations of unplugging and switching around those cables seemed to allow it to boot with the disk in but not recognize the disk.
@BonGart The largest IDE disk I can come up with at the moment is a 6 GB antique, so I'll have to try that. Perhaps I may be able to hot-swap in a SAS/SATA drive after boot from a minimal Linux machine.
@hexafraction I just figure if you are only futzing around at this point to get it booting, it's a way to build a system and then add to it as you go. But then again, I'd be booting to a LiveCD of some sort at this point
I can't really tell if it's the IO card or that backplane, but even with both SAS and the IDE links disconnected it still tries to boot a RAID kernel correctly.
@BonGart I can get live CDs fine. I just can't exactly boot from live CDs for every boot.
No disks, obviously, except the CD and the mounted squashFS root partition.
well, I have to help someone with tech at work, and work takes priority, so sorry I can't keep helping; The Comms Room (SF chat) may have additional help... just explain to them the same thing or link them to here; at least now you know you have SAS going out from the IO card (the RAID card I assume you mean) to SAS backplanes for the disks
I'm honestly never going to care about RAID, as my use for the machine will not require anything other than a new disk, a quick package reinstall, and quick reconfiguration(with the loss of maybe a day's work) should a disk fail.
@hexafraction not talking every boot (unless you customize an Ubuntu or LinuxMint build so you have a stealth System that vanishes when you remove the CD)... but if the system is attempting to boot to the blank drives you are connecting, and there's nothing to connect to, would it then boot to a CD/DVD?
A lack of bootable drive would result in a screen with an ASCII art floppy disk and drive.
Yes, I get a splash "booting controller kernel" and I can enter the setup but it only pertains to disks and turning off the controller entirely. I've tried various combos of disabling caching, and disabling other features, but none helped beyond leaving me disk-less again.
@hexafraction "pertains to disks" you don't mean you can tell it what disks it should expect to be connected? Did I miss the model number on the controller?
No, when I say that it means that all operations except a minimal setup to turn off or on caching, and the controller, require a disk and/or RAID array.
Identifying the specific hardware would assist me greatly in getting on the same page. I can call up specs and whatnot. It might be something stupidly simple like the size of the drives.
@BonGart it worked! :) sorry for not responding earlier, the XP machine was shutting down on its own repeatedly, so i couldn't get to test the new config. i think the CPU is running hot, and there's no way to turn off CPU protection in BIOS so i had to power off and patiently wait for it to cool down... probably needs better cooling
@BonGart thanks for the help! i think i understand now why it wasn't working in first place (besides my first mistake of enabling ICS on the wrong network connection), it was conflicting with 192.168.0.1 which is my router.
@Sammy once the IP range was set to something that wouldn't conflict, there would be no reason not to have it on auto
@Sammy but I am glad it worked out. Windows 7 took this specific problem you had into account, and by default, set those two parameters in the registry to 192.168.137.1
@Sammy so you weren't the only one to have that problem
@BonGart i also came to realize that, in the past, i.e. when home routers were still uncommon etc. people used wired connections, where one ethernet cable goes from maybe an ADSL modem to PC1 and then another ethernet cable goes from PC1 to PC2, and they would setup ICS on PC1 to share internet connection with PC2. so there was no conflict involved, when only two PCs were used, where one of them only had 192.168.0.1 address. but it's different in my case, because my WLAN is connected to a router
@Sammy not sure... most ADSL modems I've encountered also had router firmware, hard set to only pass out one internal IP address (so you couldn't just connect the ADSL modem to a switch or hub), but usually, their range was 192.168.2.1 not 0.1
The recording head never touches the disk media ensuring significantly less wear to the recording head and media as well as better drive protection in transit.
What as opposed to the Canyon creation method :-) "our touch head technology DIGS in and really get to your data the first time"