Archive for the ‘Internet’ Category

Debian + Puppet == <3

Tuesday, February 21st, 2012

If it’s not immediately obvious from reading any of my websites, I’m a FreeBSD guy. It’s just comfortable: I’ve been a FreeBSD user since the late 90s, and I just know my way around it. Despite using various Linux flavors here and there throughout my life, they always feel slightly alien to me.

That said, for what we do at MumbleDog, FreeBSD has proven to be a liability at times. Sometimes it feels like I’m about the only maintainer of Murmur on FreeBSD (though that’s obviously not actually the case), and there are quite a few issues that have been “well, it’s fixed on Linux” that I’ve had to go and battle with myself. It just seems like a whole lot of work, so I decided to give Linux another shot.

Playing with a few VMs for test machines, I also decided to take the opportunity to give puppet a go as well. After much messing around and documentation reading, I’ve successfully managed to set up a build box, where I build a custom package for murmur and mutter, which I then push across to an apt repository, then build and sign a “Release”, before telling puppet to install it.

I’ve just finished with it, and I haven’t actually tested it from a blank VM, but I think I’m a couple of scripts away from having a “click here to add a new Mumble server” setup.

… another wave of spam comments

Tuesday, November 15th, 2011

This blog just got bombarded by a bunch of spam comments that made it past Akismet. Comments about Zunes on a post about an older PS3 firmware, discount dental supplies on a post about my truck, and the generic “I like this idea you should write more about it” comment.

When I spam comments for my site, at least I try and write a comment that’s on-topic and adds something to the conversation. :(

Eating your Own Dog Food

Sunday, November 6th, 2011

We have a few people who “work” for MumbleDog, but mostly it’s just Sabriena and I doing everything. We’re a tiny company with big dreams of hiring our friends so we can all sit around playing video games on our downtime.

In the mean time, whenever we have to get people together to figure out some issue, Mumble’s our go-to app for internet conference calls. It’s amazing how productive you can be when you don’t have to sit around and type what you’re thinking (I type at a mediocre 95wpm according to TypeRacer, and I can assure you I think much faster than that).

New Toy: NameCoin

Monday, October 17th, 2011

Since the well of things I actually cared enough about to part with Bitcoins over appeared to dry up, I decided to go ahead and pick up some NameCoins on the BitParking exchange, and register a few NameCoin domains.

NameCoin still doesn’t work on FreeBSD – I can’t manage to build a working binary and my linux_base has too old a libg++ for it to work, so I went ahead and set it up on a Debian VM to get things going.

Once I had it up and running, it didn’t take long at all to download the blockchain (stark contrast to Bitcoin proper, which seems to take an eternity), and before long I was up and registered with a few domains. The name_firstupdate command takes a bit to actually show up in the blockchain (12 confirmations after name_new) but it doesn’t take long with merged mining blowing up.

The downside is that as yet, there’s not much in the way of usability for NameCoin domains. The only two public nameservers advertised on the Wiki both appear to be broken, the public whois service appears to be broken (despite the webserver on it advertising otherwise).

I thought about setting up my own DNS server, but so far all the scripts I can find to do it involve weird PHP hackery to scrape the entire domain space and create bind zones for each one. I don’t really think I’m prepared to descend to that level of hackery, and I suspect that’s why the public nameservers are busted.

I’m thinking about writing my own DNS server for it in Python, with a caching bind server in front of it (only public for .bit domains) but I’m not sure if I’ll get around to doing it or not. It sounds like a fun project, but I’ve got so many projects going as it is.

I’d also considered setting up a .bit registrar, simply because it seems like once I’ve done this there won’t be too much work involved in it – but I am not entirely sure how to secure it. It seems like the whole system relies on having the private keys on a box someplace, and I don’t particularly like the idea of a system where one box gets rooted and every domain you’ve ever registered can be taken.

Glorious, Glorious Internet

Thursday, September 22nd, 2011

As mentioned in a Google Plus post: “I’ve come to realize the internet is like oxygen – you don’t realize how important it is until you don’t have access to it.”

Last night, after the DSL guy came out and ended up replacing a card in our DSLAM, our DSL blinked on and off for about a half hour, before finally going down completely for 10 minutes. This caused me no small amount of rage, but when it came back up we were back on “fast” data path, with our full 6mbps downstream rate. A quick speedtest confirms where we’re at workable speeds (though by no means “broadband”, but that’s a “#firstworldproblems” scenario), so I set about setting up our permanent rate-limiting/QoS stuff.

While the DSL guy was at the central office, I pulled all the crap out of our wiring closet (though in retrospect I should have swept or vacuumed it while I was in there). I pulled out three computer cabinets, two routers (which we’re using as switches and that’s it), five DSL modems, a fax machine, and an SGI Indigo2. When the poor guy got here it was a rabbit warren of cables and stuff that almost looked like a rat king if DSL modems were rats.

So anyway, so far so good – it looks like we’re back to having a reliable service and we’ve staved off the bandwidth goons for at least a few months.

Ugh, our DSL’s been down

Wednesday, September 21st, 2011

As you may not be aware, AT&T followed through with their (I was hoping empty) threat to charge users “overages” for “excess” bandwidth usage. Never mind the fact that their service doesn’t even qualify for the international definition of “broadband”, as far as they’re concerned users are soaking up too much bandwidth already.

Since I personally think it odd that the main companies “cracking down” on “bandwidth hogs” (read: habitual streaming video users) are the ones that also offer some type of TV service, we ended up going with EarthLink for DSL service instead. The EarthLink sales guy told me that the order for DSL couldn’t go through until after we’d cancelled with AT&T – so I decided to wait on doing that until after the modem got here (in case our AT&T modem was no good for it). Well, AT&T up and cancelled our old DSL sunday night, leaving us without internet access until Tuesday.

So not only were we unable to watch any streaming video, or do anything that might spend any more bandwidth so we’re charged to the tune of $10 for 50GB (which is rough on a rather large family that all but has engravable plaques that say “Netflix: most content consumed, 2010″), but now we’re dead in the water as far as working goes.

We finally got it up and running this morning, but unfortunately it’s going slow. Really slow. 1/4 to 1/3rd the speed it’s supposed to be slow – which is crap, because we’re only a few hundred feet wire-length from the CO according to AT&T. Oh, which brings me to my next point…

… I don’t think our stint on Earthlink will be very long, because if I’m not mistaken, it looks like they simply resell AT&T out here. So I can’t imagine AT&T are going to lose a bunch of “bandwidth hogs” from paying them directly, but still keep them on their network and take it laying down. So we’ll have to wait and see what happens.

Plunge the knife, then twist it

Monday, September 12th, 2011

Blog comment spammers are getting really messed up nowadays.

I comment on blogs for the back-links to MumbleDog a lot, so I know the tricks. I try to avoid spamming, and only comment if I actually have something to add to the conversation. I don’t check PR either – either I have something to add to the conversation or not, and usually by the time I’ve figured out what to say it seems a waste to throw the comment away simply because the blog doesn’t have any pagerank. I figure it’ll all add up, and it’s entirely possible that blog may blow up one day and suddenly become really authoritative in Google’s eyes.

So anyway, even though I don’t really partake in it…I know what’s up. But this new stuff they’re doing is just scandalous. Sabriena’s got a few comments on her blog that are uhh… mean? They’ll basically say things like “all you do is whine” or something like that (and let’s be realistic, most of us who do blog purely as a form of release, a lot of what we do is whining), and you’re sitting there thinking “wow, why would you say that to a complete stranger?” Then you don’t even realize it’s a spam comment until you check the comment out, and notice that the URL/anchor text is for something like “car title loans North Carolina” and it hits you.

I theorize that they’re preying on people’s ego – you perceive that someone feels something bad about you, and you want to explain yourself. It looks vague and passive aggressive to simply write another entry without showing the comment, so you leave it up – and Bam! They’ve got their back-link.

Mucking about with Mumble Authenticators

Friday, August 26th, 2011

I really don’t care too much for the name “authenticators” as in my particular case it’s not really a good name yet for what I’m doing (there’s no authentication callbacks going on at all), but I believe it’s the commonly-accepted name for Ice-powered callback scripts. If you’ve not messed about with Murmur and Ice before, it’s immensely powerful.

Basically you can run a separate script almost like a module for your Murmur server, and have it trap various events that go on, and then the script can react to them. You can do it in your favorite language as long as it has Ice bindings – C++, Python, PHP, Ruby and Java are all supported (and I might have missed a couple). I’m doing it in Python, because while I’m pretty terrible at Python, and I’m more comfortable with PHP, realistically I’m equally terrible at PHP and I think Python is more elegant all in all (though whitespace having syntactic meaning causes me no end of rage).

We had a client approach us a while back about an idea for a site she had, and I started messing around to see whether it was doable. Using the callbacks, I’ve discovered all manner of neat stuff that’s possible. When asking for an example of setting the ACLs up in Python, a guy on IRC handed me a copy of a complete authenticator for EVE Online, which dynamically creates channels and groups based on the corporation you’re in.

We’ve been messing around with the contextual creation of channels, which would be great for a gaming community or maybe voice chat for a free online dating site (especially when coupled with an in-browser client that a few people have been working on, but thus far nothing is usable). MumbleDog (and I, personally) took ownership of a patch a while back to make callbacks for text messages, allowing you to write server-side IRC-style bots and let you do text message control over the server – again, all based off the ACLs that are in place.

I don’t know whether the client is going to go ahead with their project or not, but even if they don’t I had a lot of fun and learned a lot (and stumbled upon a few minor bugs in Murmur we can fix later on) and working on my Python skills is always a plus.

Buy Shit with those Bitcoins!

Monday, July 4th, 2011

If you’re dabbling in Bitcoin mining (or better yet, selling stuff for BTC), one of the best things you can do for the economy is turn them around and inject them back into the market by buying something from someone else. It’s no secret that at the moment the vast majority of them are simply taken to the nearest exchange and flogged off for one form of local currency or not, and that that’s going to pose a real problem for Bitcoin adoption in the future (namely, if it keeps going like this there’s a pretty good chance there will never be any widespread adoption).

I’ve sold some Steam games, and now it’s my chance to win some too. MetaCoin is gearing up to do Groupon (I think, I’ve never used Groupon) style promotions with Bitcoin merchants every day (they’ve already done some, sparsely). We were thinking about doing it for MumbleDog, but I just don’t think it’d be a good fit with the way we require a USD payment first.

Anyway, as I understand it, you “buy” whatever offer MetaCoin has up for the day, and then when the minimum number of people buy in to the group deal, the merchant works with you via a voucher/coupon or whatever and you save quite a bit.

Trading Bitcoins for other people’s shit is about the best thing you can do with them – besides holding onto them, hurrr – it means you don’t gotta deal with possibly shady exchanges, or getting ripped off by someone with Paypal. Plus you’ll have a novelty item: “check out these cool socks I bought with internet money I made myself!”

Hehehe FB

Sunday, March 6th, 2011

I think it was slag-off Facebook day on IRC or something, because everyone seemed to jump on the dogpile. I’ve long opined that sites like Myspace and Facebook are idiot herding – the apparently mental disability appeal of them is just a tractor beam for all manner of idiots with an opinion and scant to back it up.

My problem is I have a huge issue ignoring it. The XKCD comic “Duty Calls“? That’s me. I lost contact with a “friend” I’ve had for about a decade and a half, due to her chronic inability to shut the hell up on matters of religion. I don’t generally go around posting my dirty laundry with the world’s organized religions, but I’m not going to be bombarded with why yours is the greatest and everyone else is “evil” without saying something. :(