Users rage as Fasthosts virtual servers go titsup... again Webhosting company Fasthosts has been hit by a severe outage that has seen the collapse of its virtual servers. The systems fell down early this morning and even though engineers worked frantically to restore services, some users are still reporting problems. In an announcement on its website at 9.36 BST, Fasthosts said: "A …
Or rather this is what happens when you (or rather the bean counters) think you can safely outsource your data centre to a bunch (and I don't just mean FH here) whose strategy seems to be to say 'Look at us, we're really big, isn't all this stuff impressive,much bigger than you'd ever have so it must be much better. Oh, and have you seen how cheap it is? Think how much money you'd save. No, really cheap. Ever so cheap. You'd be a mug not to want to save this much money. Of course it's safe, how could it not be? And it's really cheap.'
So you outsource, things go titsup, and guess what? Instead of being able to escalate (to a single director if it gets bad), re-prioritise your recovery and so forth, you're now just one voice amongst lots, all clamouring to get those critical systems back on line, and you're probably nowhere near the biggest customer so how much influence do you really think you've got?
Economies of scale (for the provider) exist mainly when everything's going smoothly. Scaling your recovery so that you can get away with fewer bods requires investment in tools, techniques, developing robust procedures, testing... all those things, in fact, that cost. When the market's in its usual race to the bottom pricing mode, guess where the corners get cut (clue; it's not the director's bonuses). So, for those screaming at FH for not being perfect, I'd have to ask: what planning did you put in to make sure that when a big outage in your (now outsourced) data centre happens, your business didn't take a disproportionate hit? Or did you actually believe the salesman when he/she trotted out the usual soothing words? And how were the unicorns?
I'm not against outsourcing per se, despite the impression you might have got from the above. I'm against the school of thought that seems to believe that because you've outsourced, you can abdicate all responsibility as well. It's your business, not theirs.
Last time I had a VPS with Fasthosts, I was running CentOS and it was hosted on Hyper-V which appears to be what Fasthosts's infrastructure is based on.
Every couple of months it would stop responding and behave as though the drive was corrupt. I only lasted one year before moving to the rock solid EC2. Which is Linux based. Never had any problems at all since.
This post has been deleted by its author
Err... Yes. You can setup NetApp and Dell badly, or you can do it well. If the hardware is unreliable (your implication, I think?), you can mitigate against that with clustering and redundancy. If it was all homebrew hardware, I'd be surprised, but NetApp and Dell are both good, solid makes, albeit not Proliant and EMC.
"Servers that are brought back online will continue to experience degraded performance for the next few hours until the work is complete."
That'll be the RAID rebuilding, hope they haven't used large disks, as it could be days! I think it was botched upgrades to the arrays firmware last time, wonder if it is too. Maybe I should go up to Manchester and see how busy they are in the DC cages.
not with fasthosts, im not that stupid... but i do have hosting in the US (webzpro - who i have nothing but praise for)
Im now looking to move over to a UK based host, who would people suggest?
ive been to a few review sites but when i see fasthosts in the top 3, its best to avoid.
I used to host with FH (when I started I didn't know what I was doing). After several F-ups and the worst support known to mankind I moved to Linode. Now I manage my own server, it runs perfectly (currently on an uptime of over 380 days). Linode has a UK based data centre you can choose when creating your node.
I'm not a sys admin but using their guides and templates I've managed to setup a LEMP stack and postfix/dovecot email server for my girlfriends business - all runs like a dream on a tiny monthly price.
I could recommend my current hosts, but then I'd have to kill you. (i don't want more muppets using them !)
Try Redstation, they were a total bunch of cunts when it went wrong on them, no backups, no idea - told the server was dead on conference call with the chief iTard and managing director, he was a little bemused when I told him I'd downloaded the 150mb CMS site of the 'server that was totally dead as the disks had failed' via FTP but sadly I didn't do the rest as I assumed they had a fucking clue/backup. Then they decided they didn't like it.
"After much deliberation over the future of the H-Sphere Hosting platform we have reluctantly decided to withdraw the service and close it down. "
Followed by fuck off in 6 months tough shit we give up
Good Riddance
So Redstation, try them, can't be worse than Farcehosts...
This post has been deleted by its author
Just for clarity, assuming it was Virtual on Hyper (nothing to suggest that's all they use), sending a cursory pile of blame in the direction of MS sounds a little unfair if classic el Reg.
As for as I have seen the issue is not the hypervisor tech, but storage. Hyper is really very good actually, I but if tardhosts are not using it properly by ensuring storage isn't the weak point it hardly MS for the blame
What's embarrassing for Microsoft is Fasthosts are pretty well a hosting reference architecture for Hyper-V, System Centre and everything else MS do. That's not to say this is necessarily an accurate reflection of what a well run MS powered hoster could do, but given the amount of publicity MS have generated for Fasthosts, this kind of incompetency can't be welcome...
Once upon a time they were quite good if you wanted basic cheap webhosting and email. Then it became a fight to the bottom in the market and to maintain their low prices the quality of the service shot downwards.
I bailed out a number of years ago after a series of outages and giving all of the users passwords away to hackers.
I echo what Mr C Hill said.
Used to be reasonable price and reasonable service, then got cheap and crap.
Stopped using them 4 years ago after the many arguments telling them what was wrong with their servers and pointing them to various test websites to prove to them that no one outside of their building could access the servers... Got quite frustrating!
Our customers who use fasthosts or parent company 1&1 cause us no end of grief and it's always our fault not farcehosts until we always prove otherwise, indeed this month their dns issues have caused us no end of grief, still they admitted to it in the end, the power of twitter! Wasted us a lot of time and nearlty lost us some customers because fasthosts continually told them there was no issue.
How do they survive, simples, generating enough new business to replace the lost customers, offering something 'free' like 1&1 do to make up for issues, such as 2 months free hosting for a 4 hour outage, somehow that keeps enough customers happy.
I'd recommend intrahost.co.uk, used them for years, very helpful, based in the UK, UK staff, UK datacentres and nice cheap cloud servers.
Hosting isn't easy when you analyse all the resources and skills that are needed but any self-respecting hosting company should be able to get the basics right. What really matters is how you respond when things go wrong and believe me, things always go wrong.
I don't want to go all sales spammy but there are those of us (out here) who are different. We expect to be contacted by phone. We have UK only support staff who can help and more importantly, who genuinely care about getting customers back on line as soon as is humanly possible - on the very rare occasions things go horribly wrong. And we have directors who do things and who talk to customers and who sometimes have to take some flak - I know, I'm one of them.
You just need to choose your hosting partner wisely!
Trouble is, most purchasers are drawn to the big names or the big marketing campaigns; or sometimes (oft-time) to the lowest cost. He (or she) who lives by the sword, dies by the sword as the old saying goes. So my suggestion would be to work with a partner who cares and who can do this stuff well and support it even better.
Happy hosting!
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2023
