Linux not even starting ???

[linuver451]

I am getting this no matter how I format .According to me this seems to some hardware issue .This happens everytime during installation of red hat enterprise linux 5 server edition .Can anybody clarifies it and give some remedy.

http://img247.imageshack.us/img247/8351/z000op1.jpg (http://imageshack.us)

[happybunny]

look into adding noprobe=hda to the grub line for the kernel.

[bwkaz]

Well, the first line in each error (the one that mentions "{ DriveReady SeekComplete Error }") just means that the drive is signalling an error. That message comes from a set of status bits in an IDE byte somewhere, and the Error bit is turned on. The second line (the one that mentions "{ DriveStatusError BadCRC }") is reporting which specific error (or errors) is/are happening. These are error bits in an IDE byte somewhere which is only used if the Error bit in the status byte is turned on.

The BadCRC means that the CRC value (basically, an error-detection code) that the drive calculated was different from the CRC that the kernel calculated. This means that either (a) the CRC itself got corrupted between the drive and the kernel, or (b) the data got corrupted between the drive and the kernel. Devices between the drive and the kernel are the drive cables, the IDE controller, the PCI bus, the north-bridge, memory, cache, and CPU, so it could be any of those.

The most common cause in my experience is interference in the cables (but I don't have a ton of experience with these errors, either, so take that for whatever it's worth ;)). In particular, I was getting exactly this error for a while when I had my hard drives on a rounded IDE cable; switching to a flat, 80-pin ribbon cable (the one that came with one of the drives) fixed the interference, and the CRC errors went away. See if you can swap the cable for a different one (preferably a known-good one, but anything else would be worth trying).

If it's coming from a bad IDE controller, I don't think there's any good way to fix or test for that. :(

If it's coming from interference on the PCI bus, it'd be worth checking if you get weird errors from any other PCI (or AGP) device in the machine.

If it's coming from a bad north-bridge, about all you can do is replace the motherboard. :(

If it's coming from bad RAM, you could run memtest86 on this machine overnight and see if it comes up with anything.

I'd recommend running memtest86 first (ideally via their own bootable CD), and trying to swap out the cables second. After that, if you still haven't narrowed it down, see if you can get a different motherboard to test with (though I realize that's hard).

[linuver451]

Is these errors have something to do with hard drive ? And thanks for help

[leonpmu]

It could also simply be that your hard disk is failing... is this a brand new disk??

Or it could be faulty cables, my unfortunate experience has been that this is normally the eginning of the end for the hard drive..

[bwkaz]

leonpmu -- Were you getting the same flags set in the error byte? (I.e., DriveStatusError and BadCRC, but nothing else?) If you were getting different bits set when your drive failed, then it's still possible that the cause here is that the drive is dying -- but I think it'd be a lot less likely. Normally when a drive fails, it doesn't return any data at all; bad CRCs don't usually happen in that case. (At least, not in my experience...)

But running some SMART tests might not be a bad idea either, just in case. (I believe that can be done with smartctl, if you have that.)

Is these errors have something to do with hard drive ? Sort of. The error is that a CRC was wrong, but the reason the error showed up at that time wass because something tried to read data off the hard drive. It doesn't necessarily mean the drive is bad (though it might); the problem might be later in the data path. But yeah, the drive is involved.