Just built a new machine for a desktop and thought I'd give the R6 installer a try. It never gets to the LinHES background and ends in what I believe to a be a kernel panic (caps lock and scroll lock on keyboard just blink endlessly).

The installs begins to load daemons and dies after "updating module dependencies":

Hardware details:
C2D E6500
DFI Bloodiron P45-T2R motherboard
Geforce N210 (PCI-Exprtess 16x with 1 gig)
2x2 gigs of PC8500
IDE HDD (ATA100)

Here are pics of the monitor before and after it dies.




Link to flyspray task: http://linhes.org/flyspray/index.php?do ... ask_id=581

I have successfully installed Arch 2009.8 (x86_64) on this same box, except that went on a SATAII HDD. I have an old 160 gig ATA100 drive just lying around and thought it would be pretty harmless to install R6 on to it w/ this machine to test the installer. I know the hardware is good. I have stress tested it using prime95 doing large FTT's for over 24 h and no rounding errors. Memtest 86+ ran for over 18 h with no errors. I believe this error to be unique to the LinHES installer, not a function of hardware.

_________________
Retired KM user (R4 - R6.04); friend to LH users.

Graysky

Check your bios and make sure that your booting to IDE and not ACH raid. Sometimes the bios can be cranky and cause the machine to wig out. I have two different machines one likes everything to be IDE even if it is Sata and the other one is exactly the opposite. It's worth a look.

Good suggestion. The boot order option in te BIOS doesn't have an IDE/SATA option, just hardrive. I will try switching SATA from AHCI to to IDE mode and see if that helps. Stay tuned...

_________________
Retired KM user (R4 - R6.04); friend to LH users.

Problem solved, but the solution confuses me! Turns out that, it had nothing to do with the SATA mode. The R6 installer seems to be really sensitive to CPU Vcc (vcore for the CPU). As I said in the original post, this PC was not built to be a LinHES box. It was built as a primary workstation for a friend.

Anyway, it's a low-end C2D (E6500) which has been running at a mild overclock (about 18 % over the stock setting). As most do with overclocked PCs, I have been stress testing this one at this set of BIOS settings/overclock level and it shows no signs of instabilities. For example, it can run large FTT's in prime95 for >24 h, it can run memtest 86+ for >20 h, and can also run over 1 hundred iterations of linpack doing some pretty huge problem sets without a single error. Anyone who has been overclocking PCs would agree that this system is a "stable overclock." That said, Windows installs and runs just fine on it, Arch Linux installs and runs just fine on it... BUT for some reason, it can't run the R6 installer!

What was the solution? Adding just 12.5 mV to the vcc. After doing this, R6 comes up to the installer. I haven't actually installed it yet; I wanted to see if I can get my head around why the R6 installer needs that extra voltage first. Anyway, most people probably won't care about this because I think that most people (myself included) wouldn't overclock their R6 boxes. Again, this box wasn't built for R6, I just wanted to see if the built would qualify as a Tier 1 system. After I run some more stress tests on it, I'll go ahead and bump up the vcc again and see if R6 runs well on it, and if so, post the full config in the Tier 1 forum.

_________________
Retired KM user (R4 - R6.04); friend to LH users.

Does the R6 installer run when the cpu is set to recommended voltages and not overclocked? If so, then all is well.

When a system is overclocked, a single unique instruction sequence can cause excess ringing on a digital signal line or crosstalk between lines, which can cause a system error. It doesn't always have to be load or heat related.

Upping the voltage a bit may fix it. lowering the clock would be the conservative approach, but that's usually the road less traveled for the true overclocker!

Agreed. I didn't actually drop it back to the stock settings, I just added a little voltage and all seems well. Given what I've observed with the installer hanging, I am beating on the system again (using the original lower vcc) with linpack and prime95 running in the background. I also have winrar running it's continueous self test which also has error checking.

Maybe it's excess ringing or crosstalk as you pointed out. I've been overclocking computers for nearly a decade now and haven't seen this before - I can't explain it

_________________
Retired KM user (R4 - R6.04); friend to LH users.