BOT House|RELOADED - Part I

At long last, BOT House is finally reincarnated.  And on a much perkier Intel box with a newer, 64-bit version of Debian.  The new, official name is BOT House|RELOADED or BH|R for short.

So much hardware has crashed and burned this year it's hard to keep teack of it all.  First, it was the proxy project box.  It died of—what else—hard drive failure.  It was a strange setup in the first place: a (hardware) "RAID Nuthin" array spread across an IDE and a SATA drive.

And to complete the nightmare... on LVM.

But there were backups.  I re-installed everything on an external USB/IDE drive temporarily just to keep things running.  Later I bought a pair of 3.5T SATA drives & a new RAID card.  I mirrored the drives and plan to use it as the main backup for all this crap I shit out.

But before I finally got around to taking it off the USB, UPS's started to shit themselves.  Power failures have been brutal this year.  It used to be all I had to worry about were a few minor brown-outs during the beginning of "air conditioner season", but this year multi-day blackouts were far too common for my comfort level.  Two UPS's died.  I replaced the batteries in one and upgraded another from 350VA to 1000VA.

A few weeks after the derecho hit and knocked us out for four fucking days (two off, one on, two more off) it was getting a little windy outside, so I decided to check the Weather Channel for a forecast.  I turned on the TV, tuned in, and no sooner than they said "... high winds approaching our area..." the entire house went dark and stayed that way for another forty-eight hours.

A couple of weeks later, there I was, minding my own business and limping away on the USB drive in the proxy project box when one day, after mowing the lawn, I sat down and searched for images of Mossberg shotguns (for this story) on Google.

And... nothing happened.

Then I hear this "click click" sound coming from BOT House.  I switch to the console.  The last thing I saw was a message that said "Replace UPS battery" before I tried to reboot it.

It didn't reboot.  It just went click click click...

I spent the rest of that afternoon recreating the router & firewall on a bootable USB version of the Backtrack5 LiveCD and ran that for a couple of months before buying all the new hardware—computers, UPS's, hard drives—for everything and re-engineering the whole DinkNet NOC from the bottom up.

The things I do for you kids!

Running Chromium as Root on BT5R2

Well, sort of.

You can't do it.  They (the omnipotent and wise developers of chromium) won't let you.  If you try to run it as root, you'll get this...

You can run it su'd as a regular user while logged in as root, but it takes a little X-Fu to get it done right.  First, create your user with the "adduser" command.  Then, give him access to the X display with the "xhost" command.

Here, our user is called "bob":

xhost +SI:localuser:bob

Now start a terminal session and "su bob".  Then run...

chromium-browser --user-data-dir=/home/bob

...and you're in.

This is not unique to Chromium,  The Tor Browser Bundle for Linux will also not let you run as root, which is a bitch and a half.  This trick will work with Tor as well, but it will bitch about the user's profile.  So far I haven't found (or looked for) an equivalent "user-data-dir" switch.

Chromium isn't in BT5R2 by default, so you need to install it with...

apt-get install chromium-browser

Neither is the Tor Browser Bundle, but I'm sure you can figure out how to install that.  It's as simple as downloading & extracting it.

So anyway, BT5R2 is the best version yet, but as usual I have my complaints.  I won't bore you with them now.

Give me another week.

One More Time

Well, those were all great ideas. Or so I thought.

BH still hung, with the same symptoms. Back to the legendary drawing board.

Since time immemorial (the 90s), when I had a stack of 10mB NE2000 NIC clones, I learned a Wonderful Thing about Linux and NIC driver modules, specifically, of course, NE2000 NICs.

If you had a system with two NE2000s, without fail the kernel would only detect the first one at boot time (and, yes I tried boot time parameters). If you had one NE2000 and one SMC NIC, the kernel would detect them both. I eventually found that if you compiled the NE2000 driver into the kernel that both NICs would be detected. So that's what I did.

Time went by and the stack of NE2000s went into the landfill to be replaced by a stack of RealTek 100mB NICs (RTL8139s). I found the same issue and ran across another: not all RTL8139s were created equal. Some liked one driver (8139cp) and some liked another (8139too). I ended up building both drivers into the kernel because for one thing, you can't tell which NIC will like which driver just by looking at them (I have at least one, now marked with a Big Black "X", that doesn't like either).

This may not have been a good idea, although once again this is exactly how EXPIV is set up, except EXPIV has only one NIC.

So tonight I rebuilt the kernel with only the 8139too driver, which seems to be the preferred way to go from what I've been reading on the Web.

I am not optimistic, but one of the symptoms appears to be that the CPU is spending 100% of its cycles on IO when the system hangs, at least according to the Gnome panel system applet that runs on the desktop. That, combined with the odd network behavior BH shows when it's hung, leads me to believe the drivers may be confused about which one is in charge.

Maybe I'm grasping at straws at this point, but time will tell.

Experimenting with EXP IV

Last week I bought another old PC from the junk computer store, just like the one on the left there (except mine has a CD drive).

What a piece of CRAP, eh?

I always liked those old iPaqs. Back around the turn of the century, they were pretty futuristic. I have a fond place in my heart for them, and since the price was a mere $39.95, I figured what the heck, take it home and have some fun.

It came with Windows 2000 Professional, so I wiped the drive and installed Xubuntu 9.04 on it. Runs Xubu like a top. 1gHz PIII & 512MB of RAM with a 20G hard drive. It's not going to blow anybody's doors off, but it's a nice little "Mail & Web" PC.

But I have one of those already so I decided to play around with it.

I moved a copy of EXP IV over to it, called it "EXP V", made sure it wasn't advertising itself on the Web, and fired it up just for funsies.

And the little bitch pissed me off! It played UT much better than my Hot Shit AMD64x2 with four times the RAM and processing power.

So now I'm mad and I need to know WTF is the deal with the AMD box. I'm not sure where to start but I think a good place to begin is the NIC, although I've been down that road before.

But it BUGS THE HELL out of me. Something has to be done.

Screwing The Pooch, Linux-Style

My dear wife, Pinky Dink, has been running Ubuntu 6.04 LTS (LTS="Long Term Support") for close to three years now. She has been quite pleased with it but I've been bugging her about upgrading to 8.04 LTS for months.

But she's seen how my "upgrades" go (the long saga of EXP IV being the most recent) and wanted nothing to do with it. So we compromised for a while. 6.04 came stock with FireFox 1.x, which is dreadfully out of date. I found a hack on the Web and upgraded here to 2.x, which is only woefully out of date. The problem there is the upgrade was out-of-band, so she stopped getting updates to FireFox or important plug-ins like Flash (which is full of vulnerabilities and has been hacked to death lately).

Not a good situation.

So I finally talked her into it and she let me do it.

I anticipated the SCSI issue. Maybe you haven't noticed, but IDE hard drives are obsolete. SATA is now king and to simplify things, Linux now treats all drives as SCSI drives. If you upgrade to a new kernel and your root drive is on /dev/hda1 (the old IDE drive designation), the kernel will choke on boot because it thinks it should be /dev/sda1 (SCSI). The way around that is to either use LVM (Logical Volume Manager - which I despise) or to label your drives before you upgrade, which is exactly what I did, having learned that the hard way over several VM Linux upgrades and the EXP III/IV debacle. Use e2label for ext2/3/4 drives and mkswap /dev/whatever -L yourlabel for the swap partition. Then edit /etc/fstab with LABEL=yourlabel where /dev/hdax used to be. Reboot to make sure it works. If it doesn't... call me. Leave a message.

I knew there were going to be other issues. VNC4 is a wash, as I mentioned before. I was prepared for that. The alternative has always been Cygwin/X, which is slower and suffers from a lack of a dependable clipboard, among other things.

What I didn't expect was problems with her wireless drivers. When I first set that box up back in '06, the RaLink drivers had to be built from alpha or beta source and had to be rebuilt every time there was a kernel upgrade.

But over the years, the RaLink drivers were mainlined into the Linux kernel and that problem disappeared.

Turns out, it's back. With a vengeance. And it's apparently due to some "philosophical" differences the developers at Ubuntu have with the driver. For details, see this exasperated posting from some poor schmuck who can't get Ubuntu to admit they have a problem. I followed his lead, downloaded the source code, and blacklisted the stock Ubuntu drivers. Wireless is fine again, but I will be fucked on the next kernel upgrade.

One problem down.

The next was the remoting issue. Pinky's box is in the kitchen, but I always used VNC4 to hook up to it and do patching and housekeeping (which is a really bad idea - if your session gets trashed during an upgrade you can leave the box in an "undefined" state - but I'm prepared to take that risk because I'm too lazy to get up out of my chair).

As I said, I expected to fall back on Cygwin/X. Silly me.

Cygwin, for those who may not know, is a set of Linux utilities that run on top of Windows. Cygwin/X is an X server that allows you to run graphical programs or full graphical desktop sessions on another Unix box across the network while you're logged in to Windows.

This is very nice, considering it's free when most third party X servers (such as Hummingbird Exceed) cost hundreds of dollars.

The problem is, with all due respect to the Cygwin developers, they keep fucking the god damned thing up.

Upgrading Cygwin is a crap shoot. Something is always guaranteed to be broken.

So when I tried to connect to Pinkie's box and the X screen kept bouncing, blinking on and off, I knew I was in for more fun. I crossed my fingers, girded my loins, and upgraded Cygwin.

It worked. Or perhaps I should say I simply haven't found what's broken in it yet.

Now I have the box where I want it to be (almost). In fact I'm posting from it now, wirelessly and over a network Cygwin/X session! I'm 90% pleased with the way things turned out and this is still the best Linux box I ever built.