number 9

I’m sat here listening to Frank Black s first (and best) album for the first time in almost a decade, and browsing New Scientist.

They’ve got one of those hideous Flash banner ad trainwrecks that makes you want to browse it in lynx. Seriously, I can’t even see half the site from Ubuntu with Firefox. One thing I did just see was a gooooooogle ad for The Surface of The Sun. The link text says ‘the suns surface is made of metal and conducts electricity!’.

I followed the links, and the images are from the SOHO observation satellite, and are probably pukka. I mean, red giants synthesize iron don’t they, and ’metal’ is a pretty wide category.
My bullshit detector is on the blink at the minute, and I don’t know whether newscientist get to vet the ads gooooooooogle put on their site. For all I know its’ another ‘Time Cube’ site. (Considering I have a degree in this stuff, that’s a bit worrying. Like I said, I’m a bit short of sleep this month.)

What? The point? Oh yeah, sorry. We were at a barbecue with my 2 year old over the weekend and she and her mate were playing a game which involved catching falling stars and eating them. I asked her what they tasted like, and she said ‘metal’. Just like Jesus .

Oooh! I love this one!

Paaaarry the wind – high, loooow….

Just a quick note to self : don’t waste any more time on Bluetooth.

Brief history:

2002: initial impression:

1. Good name
2. A bit like cables, only
   • expensive
   • brittle
   • gives you tumours
3. rubbish *NIX support

2003: on-call bluetooth phone and laptop.

1. buy BT card for Zaurus so I can ssh in and reboot VBCs shit servers from the gutter outside the pub. Waste of time. Sadly turns out
   • GPRS Bad. overly complex way to pay through the nose for 90s era bandwidth.
   • BT Linux support pretty bad.
   • (Open)Zaurus bluez support so bad I sell the little bastard because I am tired of fighting the urge to throw it against a wall.
2. OBEX not that awful, strangely. Load ringtones without calling 0898 numbers. (Wait. OBEX is part of irDA, not BT. That explains it.)

2004-2005:

1. Security issues (viruses, piss poor crypto) makes WEP seem merely dumb by comparison.
2. bluetooth headsets make you look like a plum. With schizophrenia.

about 2 months ago – present :

1. swathe of articles turn up on BT applications.
2. Low power / PAN very attractive from a ‘teeny computer’ point of view
3. get the spec out of the library
4. start looking around ebay for ‘dongles’

But it’s

• pricey
• complex
• still stupid ( and the SIG seems oblivious to these issues – at least the 802.11 guys held their hands up to WEPs holes).

Think I was bewitched by BT as an interface to devices with no UI.
But if I have mains power, I am probably next to a PC. And I live in Wales.
USBnet is just a better ( e.g. securer, faster, less power, OS agnostic) connection than GPRS over bluetooth for 99% of uses. That might tie you to a keyboard/mouse interface, but most mobile/PDAs are as pleasant to use as a lump of hot lead smeared with dog poo (for my RSI blighted wrists anyway).

You know what? Screw bluetooth. If I want wirefree, 802.11 has worked fine for almost a decade..

We don’t need another solution in search of a problem. Its ‘killer apps’ are either irrelevant in the broadband age or never made sense in the first place.

Take the ‘laptop in the boot’ usage profile. Thieves now use this ‘feature’ to find, er, laptops in boots. And not to check their email.
The ‘fix’ is the usual bluetoooth fix- if you want to secure it, don’t use bluetooth. Thanks.

I hereby predict that soon you’ll be walking down a dark alley with a mate and behoodied ne’er-do-wells will emerge from the shadows. One will produce a phone and press a button.
Whichever of you beeps has the expensive phone.

Bagsy not me.

About 18 months ago I worked at $VERY_BIG_ISP, whose bazillion legacy servers were interdependant in hard to spot and baffling ways. Changing foousers quota might cause an outage on a server you’d never heard of which scped files to ~foouser so a third server could pick them up from behind a firewall.

To help, they made a very nasty and bureaucratic change management system. It didn’t help (and made you choose the solution with the least red tape rather than the sanest one – see above).

Anyway, it couldn’t answer things like ’What cronjobs around the network rely on the location of a particular CGI script?’ or ’I am in single user mode and want the webserver up. What other services (database, DNS) need to be up?’. Admittedly you could stumble blindly around for the answer to these questions, but 20% of our work was clearing up the shit that had occurred when we guessed wrong (75% was wasted on the change management system, which didn’t leave much time for jobserve).

They needed to describe a tangle of components around the network with nothing much in common ( language / packaging system / OS) except an interdependency.

My basic idea was to ‘draw’ a graph describing how all this ”complexitti spaghetti” was tangled up. The nodes could be services like ’NTP service’, ’20Gb of RAID5’, ’accounts database’, ’IMAP mailbox for user@visp.com’, or whatever. Links would be dependencies,

You be able to show that to be functional, ”user@visp.coms imap mailbox” depended on cyrus, working DNS, LDAP for authentication, and so on.

With this Golden Hammer:

• system documentation would be a graph
• system deployment would be a graph you fed into a tool
• the current system state could be checked against the graph
• changes would be described in terms of updates to the graph
• system initialisation and installation would be a graph

In this happy wonderland, you’d just cross out ‘courier’, write ‘cyrus’ , hand the sketch to your robot minion and go kite flying. (Of course, some poor sod would have to sit down and document the shambles in ‘graphese’ in the first place. Exercise for the reader.)

Despite your obvious skepticism, I actually started implementing this pipedream. It was called ’ratking’ and was really just an excuse to keep my hand in with Ruby. Plus I liked the name.

After a few commits I’d reimplemented rake. Badly.
Each node was essentially just a Proc and a list of dependencies.

I had a nice test suite though.

Then I found that you could’nt serialize Procs, got a new job somewhere less doomed, and gave up on it.

I got back from holiday yesterday to find Jamis Buck has written SwitchTower, which is (at a quick glance) network rake. I’ve just started looking after a maze of twisty webservers, all alike, so his timing couldn’t be better.

Still think my name was better though.

Finally made it to UKUUG at Swansea last Saturday.

The main event for me was Xen. (Admittedly I was sat in the Taliesin ‘cluster roadshow’ all day). It was good hearing from users (Marcus) alongside the Xen team (Ian), to be reassured that it wasn’t hype. They both humoured me when I asked my questions, which was nice of them.

I somehow missed VM migration in Xens feature list (moving entire VMs so kernel data structures move too, so TCP sessions are preserved) – up till now I’ve thought of Xen as a ‘fast VMware’ rather than a natural way to package services for resiting around the network. Another use would be snapshotting a crashing system (of which I have plenty ) and putting the image on ice to debug once the immediate crisis is over – Ian mentioned VM breakpoints being added post 3.0, which would really help here.

Once Vanderpool arrives then there’s also the potential for running linux (or better yet, plan9) like Seti@home as a low priority background VM on public access Windows terminals.

I got the impression there were some issues with running Xen images off an NFS share (loopback locking, etc.), which would mean you’d need a SAN to get the full benefit of it (then you get disk snapshots, which you need for the debug stuff). I’ll run diskless clients in my VMs at home if II ever fancy playing with it.

Inevitably, all the other speakers were compared to Xen.
Poor OpenMosix looked disappointing by comparison: no security, limited platform support, and ( the main showstopper in my book ) the fact that the ‘home node’ of any process is a single point of failure for it, however many nodes you have.

The IBM guy (they have hypervisors in silicon, so it’s old hat to them) was doubtful about how robust Xen is. I suspect the price and GPL factors will make all the difference to its progress.

Anyway, the Powerpoint slides are on their Wiki , have a look.

As an added bonus to the day, Oreilly had a stall. So I got my Rails on faster than if I’d waited for Amazon, at a lower price. The place was crawling with Pythons but there seemed to be plenty of people thumbing through it.

Despite the German guy who seemed to spend all day shouting ‘use plone!’ at interested browsers

loving typo 2.5.

Especially loving migrations
( They don’t work with sqlite yet, so I had to convert the old db to mysql before being able to use them, but nothing too scary)

Interestingly, most of the goodness is in the admin screens, rather than the public parts? ‘By Railers, for Railers’, I suppose

Still some minor niggles with mod_proxy and how it constructs absolute urls to the wrong place, but that’s probably fixable after some sleep.

( Surely you’re joking, Mr. Feynman? just finished downloading from Audible, so I’m off to bed. Assuming the baby has got over her insomnia, I should get the kinks out of the blog tomorrow.)