Building a linux server
Saturday, November 5th, 2005
Sigh. I was really hoping to have nothing of interest to say about the process of upgrading my server hardware.
After putting everything together, I’ve come across an intermittent. The machine boots up, I can fiddle with the BIOS (not much of interest in there), and boot from a CD. Cool. That’s the hardest part behind me. But at some random point during an attempted Linux install, everything locks up. No errors. No hints.
It is all brand new, so I can’t just try the trick of “what was the last thing you changed before it broke?” Instead, I’m stuck with a strategy of “replace one thing at a time until it works, or all parts are replace, or you run out of money”.
The part that I suspect the most is the SATA controller. A reasonably authoritative-looking web page has this to say about my nVidia SATA controller:
(link) Nvidia NForce3 Go120, Go150, 150, 250, 250Gb, and 150 Pro (SATA-I); NForce 4 MCP, SLI, Ultra (SATA-II); and NForce Pro 2200 MCP and 2050 (SATA-II) chips — fakeraid. This motherboard chipset family uses yet another type of fakeraid called nvRAID. Beta libata driver “sata_nv” exists as of 2004-07-08 development code, included in kernel 2.6.8. Alternatively, proprietary Linux i386 binary (ataraid?) drivers, including support for the proprietary RAID functionality, are available, but, as usual, you’re better off using Linux’s own open-source “md” software-RAID driver for the RAID functionality. Note that Garzik is quoted (1, 2) as saying that “Unfortunately, Nvidia is the only SATA hardware vendor that chooses not to give me any hardware information.” (Nvidia hardware seems mainly aimed at Windows gamers, and historically has tended strongly to be proprietary.)
So, I ordered the cheapest, smallest ATA drive I could get from Newegg ($50) and when it arrives I’ll embark on my replacement strategy. In the mean time, I remembered I have a crappy old ATA drive in the closet. But it is so much easier to order a new one, than it is to look through that closet. Still, I’m a desperate man, so I dug it out and tried it. I/O errors galore… not a very good test. So I wait for a new drive to arrive.