1and1 .htaccess failures

I have a 1and1 hosting account, on their “shared Linux” platform. The issue I had was that I could not get the server to handle a .htaccess file.

What would happen is that I’d get a “Server Error 500” on something as simple as:

order deny, allow
deny from all

1and1 tech support was little help. I was told my syntax was wrong, and sent an email explaining .htaccess syntax to me. Classy.

What turned out to be the issue was my FTP upload. I was uploading in “Auto” mode through FileZilla, which defaults to ASCII for .htaccess. Which made .htaccess end up with Unix linefeeds on the 1and1 server. That seems perfectly rational for a Linux hosting package – alas, the 1and1 server doesn’t agree.

What I needed to do was force the file to have Windows linefeeds, which I did in Notepad++, then change the upload type to BINARY and upload .htaccess.

And that did it – my .htaccess files are now working.

Recovering from BIOS failure on Gigabyte GA-EP45-UD3P

After installing a new motherboard – a GA-EP45-UD3P – and successfully booting into Vista without issue, I got foolhardy and ran Gigabyte’s @BIOS update utility – for no good reason but the sheer heck of it, to upgrade from “F6” to “F7” BIOS.

@BIOS runs in Windows. I really should’ve known better. It completely hosed my BIOS, to the point where all I would see upon boot was

Award Bootblock BIOS 1.0

scanning bios image in hard drive

And a little bit more text that I did not bother to write down. I did manage to recover from this.

The TL;DR version is: Do not clear CMOS. Use a SATA CDROM with the BIOS image on it. [ Edit 2012-12-3: Or a USB HD with the BIOS image on it ]

The longer version, and lessons learned:

– @BIOS is the pits. The likelihood of rendering your machine a doorstop is high. Don’t use it; flash from within the BIOS itself if you have to flash.

– Gigabyte’s “Dual BIOS” on these boards is “virtual”. That means it tries to write a BIOS image to the hard drive, and recover from that. According to various tech forums, this recovery process usually fails. Maybe it’s the backup that fails in the first place, who knows. I could not find any information on where the BIOS image is kept on HD. Gigabyte documents the BIOS recovery process.

– Gigabyte removed the usual Award BIOS “boot from floppy” recovery routine – so the “easy” recovery of creating a boot floppy is out

– Every single thread I could find on this issue ended with “and I had to RMA the board” or “and we never fixed it”

I managed to recover by grabbing the F7 BIOS off the Gigabyte site, burning it to CD, and connecting a SATA CD-ROM to my machine. I had disconnected my hard drives at this point. This did recover my machine – to a BIOS that calls itself “F6”. I do not profess to understand that, and will not question my lucky star. I thought I had to RMA, which would have been Not Pleasant. At this point, I am going to leave well enough alone – there really is nothing in “F7” that I want or need, and for all I know, I have an F7 that calls itself F6. Unless there is a miraculous BIOS recovery mechanism that kicks in when a SATA CDROM is connected but doesn’t actually read from CDROM, which sounds more like black magic than IT to me.

[Update 2009-03-13]

I did have one more thought on the “black magic”, as I upgraded to F7 through the BIOS proper by now after all, and it does call itself F7. It’s possible that there was a copy of the old BIOS in CMOS, but that for reasons only known to GigaByte, that backup is only brought in if there is a SATA CD-ROM with a BIOS image on it present. If that’s the case, it would explain why so many people on the forums who, as a first step, cleared their CMOS, were never able to recover and had to RMA. So don’t clear your CMOS if you have this issue, just in case the backup BIOS is kept in there. Clearing CMOS won’t help anyway: This is not an issue with corrupted BIOS settings, it’s an issue with a corrupted BIOS image.

[Update 2012-12-03]

People in comments had success by copying the BIOS files to the root of a “USB hard drive”. I’m assuming that’s an actual hard drive in a USB enclosure.

Also, one person who cleared CMOS was not able to recover, giving weight to the idea that CMOS needs to remain intact for this recovery to work.