A2 Hosting finds 'restore' the hardest word as Windows outage slips into May The great A2 Hosting Windows TITSUP* has entered its second week as the company continues to struggle to recover from a security breach that forced its System Operations team to shut down all its Windows services. To recap, things went south on 23 April as malware spread over the company's Windows operation, causing a problem …
.......what do you expect for ~ £3.60 a month ?
If people insist on running their life blood and income source on something that costs so little, why expect so much when it goes wrong ?
i feel very sorry for everyone who's lost data, ive been there and its a total nightmare especially having to explain to customers its just "gone".
do yourself a favour and either pay for SLA backed managed hosting or run your own server and backup yourself - you only have yourself to blame then :)
You can easily host on Google Cloud for $3.88/month (f1-micro) according to https://cloud.google.com/products/calculator/
I'm sure AWS has similarly cheap offerings, but in my experience GCP is significantly cheaper eg. faster systems for the same price. Not sure why you'd use anything else.
It always amazes me that people "lose money" over a system that - for way less than even a day's outage money - could be doubled in resiliency by putting some inactive server in a disparate cloud, or on their own site, or whatever.
Even a VM replica, syncing every 15 minutes or so but otherwise idle, could have been spun up quickly even if it wasn't the fastest server / connection that it was sitting on.
I did a calculation once - my employer's entire systems could be stored and booted up on a single laptop. Not claiming it would be "fast", but it would be up and you could carry on business in a reasonable manner.
Literally, for the cost of a laptop with a copy of Windows Server, VMWare or whatever, I could have an entire on-site warm spare of everything we run. Built-in UPS too! When you then get into "we need a real system", then it's a cinch to spin up instances in a handful of separate cloud servers and have the same.
But, these people are complaining about "losing money" when they could most likely have just turned on a cheap laptop and got access to all their data, a recent backup of everything they had (e.g. web transactions) and maybe even got a chance to operate their services from anywhere in the world (sure, it'd take a little tweaking of settings, but at least you'd have the option).
If it'll go into a cheap laptop, since the main issue is just the storage, a simple external HDD and an automated backup routine would probably be all they need. Same as a SOHO user would do. If it doesn't have to backup fast, or retrieve data in real time, that's more than enough, surely.
I run a website and I keep a mirror of its contents on several machines at home and one on a backed up account at a local college. I update the website by updating one of my home systems and copying the changed files to the website provider. Once a previous provider lost all of the website and when they got the site back in operation I just reloaded the site. This is a site with static pages so it would be more complicated with a data back end but surely not that much harder.
I talked with a Paychex IT person who said that they have 6 duplicate versions of their data with UPS backup etc. Even with a city wide power failure of a week all of their servers continued to function perfectly.
If data are your crown jewels you should protect them.
https://www.a2hosting.com/blog/security-breach-plan/.. and
"here at A2 Hosting we provide free Secure Sockets Layer (SSL) certificates, regular site backups"
Mind you, "regular" could mean once every 6 months, I suppose.
Regular does not = frequent.
But they are frequently confused.
When people moan that some of us are too aggressive about their misuse of words it's examples like this that should pull us back.
A promise that something is regular could mean that it gets done once a decade, on the third Thursday in July.
But frequent should mean it gets done often. How often is a whole different kettle of ball games.
In either case it means very little unless quantified.
However, let those who have never skipped a backup throw the first stone. Or, um, fling the malware into the data centre.
Beautiful.
But I digress. "The Cloud" is just another word for another man's purdy compootah. And it is up to you to ensure that backups are made.
Unless it is written into the contract, treat the case as if no backups exists, and you need to make these.
"Unless it is written into the contract, treat the case as if no backups exists"
If you make your own backups you should at least occasionally check that these are what you need to restore (it might be on the backup medium but it doesn't help if your restore process has problems). If you haven't successfully tested them don't assume you have backups.
How do you get to test your supplier's backups?
Even if it is written in the contract, why would you not make your own backups too? The bottom line is that no contract or SLA refunds are going to get your customers back (and most of them are up to 100% of the monthly cost or something, so aren't worth toffee).
Data doesn't exist unless it is in at least 3 places (one of those needs to be on a different site to the other two).
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
