New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
High availability isn't about a single provider or even two.
Redundancy is N+1 and that equals THREE at minimum.
For monitoring you want multipoint to the single monitored point.
Reason is different networks could experience network blips, allows you to measure latency indicators, etc.
I'd pickup 3-5 tiny years virtual servers distributed geographically. Four to five at far point of the US and one over in Europe. That's to start
For the OP, I'm assuming HA has to be built on the application side to support this kind of setup?
So say, I have a dozen of virtual servers distributed geographically, what is the best way to setup a high-availability front-facing website? HAProxy?
CloudVPS.com
I don't disagree, and that's why I said I wasn't sure what Damian's definition of high availability was. Personally, I would never consider one system with one provider, no matter what infrastructure they have, to be true HA. Higher availability, sure, but not high availability.
I'm in AMS.
If you still are looking for someone to fulfill your needs, let me know.
@Damian we label ours as cloud. But really it is a buzz word. Even most HA providers have the same problem, the actual power to the rack is not redundant, which is why we had our latest downtime.
I use yisp (http://www.yisp.nl/cloud-vps/). They are pretty stable but as some others here said still HA limited to one datacenter as potential point of failure.
Specifically HA servers: http://www.cloudvps.com
Lowest plan is 1GB at 10 EUR/Mo, so pushing the LEB mark a bit.
I have a friend who's had one of these for about 6 months, no issues so far.
@Andre - it just looks like onapp on a san?
Unrelated - do you have a portfolio? Need a few heros/marketing/etc. designs created for a customer.
@Damian Why not get in touch with @prometeus, im sure he can sort you something HA out.
I decided to go with the make-the-app-tolerant approach, instead of the throw-money-at-keeping-it-up approach. Will be cheaper, and probably end up working better (since the app would expect that the possibility for downtime exists, instead of assuming it's always on)
We have some LEB plans on our HA infrastructure - now with SSD hardware based auto tiering
^ This