DotBlag

So, back in 05, I ordered a copy of The Ultimate Hitchhikers Guide to the Galaxy from Amazon. The book arrived. Sometime later I was reading it and discovered that, lo and behold, it was missing pages 529-560. After some consternation and thought I realized indeed the pages were gone. I wrote Del Rey books, to see if they’d replace it. Today I get this, headers included to show you it sure as heck wasn’t stuck on my end:

Seriously. You’d think after all these years there’d be better tools than BGP and better tools for working with BGP. But no. *sigh* deactivate neighbor 216.x.y.z commit

The Cast: Linux, played by a happy little penguin. OpenAFS, played by a herd of longhorn cattle chroot, played by a yak ln, played by another yak The Scene: OpenAFS running on any server platform, probably Linux. Several AFS Volumes exist across multiple servers. With many clients. Act The First. The Penguin acts friendly but is very deceiving. It pretends to be friendly and clean. But when you look really close you see it hasn’t had a bath in years.

Seriously. WTF? (click the full article to see the screen capture)

[ part two of I Cee emM Pee you ] If a woman ever tells a man size doesn’t matter, she’s certainly not being truthful. Similarly anyone talking about packet sizes is also spinning a yarn. Tunnels can and will reduce your MTU. A lot of consumer NAT devices don’t handle/pass along the ICMP Unreachable codes packet-too-big and cant fragment correctly. This means people end up ‘broken’ trying to access your site.

An open invitation to D-Link, NetGear, etc. DISABLE ANYTHING IN YOUR DAMNED ROUTERS THAT MIGHT FILTER ICMP UNREACHABLEs. We’ve been seeing occasional problems with a tunnel “somewhere” on the internet getting into our path. It shouldn’t be a problem except that it seems a lot of firewalls still filter ICMP Unreachable’s at the least. DON’T DO THAT.

OK so if you try to pronounce it incorrectly people might think you have a cold. The good news is though that Server Name Identification AKA TLS-SNI will likely be making it into Debian Lenny. It’s already in Apache 2.2.x upstream but there’s no release with it quite yet. Now we can do SSL Names Based VHosts….sorta. Only IE7 on Vista, Opera 8, and Firefox 2+ support it (sni.velox.ch is a test site FYI).

Seriously? WTF? Just had this happen when trying to click on See Larger Image of a book cover in Amazon. I have to create a new category for this .WTFMate? Click on the image to get a better look.

Yahoo! Instant Messaging recently (from my perspective) had a nearly 2 hour outage. This has brought back a problem we’ve had here a number of times in that we – de facto – standardized on YIM for internal messaging. I’m now trying to encourage everyone (again) to use IRC on our (sort of private) IRC Server. YIM also has an awesome quirk, that has gotten better and worse, of rather apparently randomly filtering messages, especially messages with URLs according to some scheme only they seem to know.

For all the retards out there who design news site software, get rid of the damned pagination in your articles. It severely pisses me off and makes it much harder to read your articles. It also makes NO sense on the web! Web browsers have scroll bars for a REASON. Lists of articles YES go ahead and paginate those, but offer the ability to change the #/page! </rant>

An interesting default change, coupled with a goof on our part caused suPHP to stop working on one of our customer’s servers. They removed the hardcoded handler – x-httpd-php, in favor of a suPHP_AddHandler directive. Compound failure on our part though, we removed the suphp.conf file. So if you’re upgrading from Debian 3.X to 4.0 or older suPHP to newer suPHP, check to make sure you’re putting in that suPHP_AddHandler! Debian will do it for you, as long as you Do The Right Thing(tm) and override rather than ovewrite the mods-available/suphp.

We’ve been seeing some reachability issues either inside Tiscali International or Global Crossing. And the finger pointing has started between them. AOL/ATDN is also affected. All of these issues are appearing to European customers. Global Crossing gets a big BOO for not having any way for me to contact them (an email to their NOC gets a “login to our portal to open a ticket or include the ticket number in your email, this mail has been dev nulled” autoresponse).

I was sick for most of last week so wasn’t really able to post anything. Rest assured though Murphy was not out sick, and thus, neither was I completely.

It always amazes me how slow mass media is, even mass media targeted at tech’s. Network World today posted this article on spammers using auto-responders to get their crap out. I don’t know about anyone else, but I sure as heck have seen that before.

Yahoo! is getting to be very aggressive against spam. Problem is we can’t deliver ANYTHING for about the past 6 days because we can’t get a dialog with them as to 1) why they’re unblocking and 2) they won’t unblock. I think ultimately here we’ll be forced to discontinue most or all forwarding services if Yahoo!, etc, continue the current extremely unfriendly-unable-to-cooperate behavior. What would I like to see? Well…It’d be nice if they used a SpamCop like process to find the ACTUAL source of an email and inform some sort of blacklist of that source, that other places can query.

rsync -arvx –progress –delete root@host:/ /mnt/target edit hosts, interfaces, and half a dozen or a dozen other configs install boot loaders reboot. twiddle… why the hell is clamav bitching about ownership… and postfix too? wtf is going on. *lights come on* AWW CRAP. GUH. UID’s *ALL* messed up. Remember, tar, and rsync, both try to do some Stupid Tricks with UID’s and GID’s. So don’t forget –numeric-ids Now, I’m leaving you all while I go work on some find … chown foo.

Or the stateful-firewall. Interesting little thing to get bit by. But to understand it at all I have to go into a little bit of explanation. We use LVS in what’s known as a Direct Routing (LVS/DR) configuration. We have a (large) number of VIP’s that the load balancers handle. The VIP’s are not on any subnet but are rather routed to the load balancers via OSPF. This makes for a REALLY confused stateful firewall when connections originate on the same subnet as the real machines handling the VIPs because they respond directly to the source, bypassing the firewall.

Looks like the <pre> tag I put in this post didn’t cause the layout to expand like i thought it would. Oh well. Looks like yet another case of mis-formatting. I’d look into it, but I’m really not an HTML/CSS guy so I’d barely know where to start.

OK If you’re a hardware designer, or embedded systems engineer/designer, make sure when you design your flash process to be customer recoverable. Bad flash’s happen. Also, blade manufacturers, other activity on the management module, shouldn’t interrupt a flash process. Looks like I’ll be returning an Intel SBXD132 due to a fragged SMP (System Management Processor — Embedded management module) that had a failure to flash. It’s just silly these days to not include the little bit of extra flash so it can come up in *some* mode that the customer can get firmware onto it.

Some days it’d be nice to be a cockroach, life would be simpler that’s for sure. So in a previous post I mentioned the trials of getting an ARC-1280ML working with 16+ drives in JBOD. We’ve decided to go with a pair of the same controller, less ports the ARC-1231ML. Hopefully they’ll fix the limitation at some point. So we’re waiting on that new hardware. Keep in mind there’s nothing really *wrong* with the ARC-1280ML, it’s just not going to do what we need it to do.