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.
The provider with the best uptime?
Does anyone have some open tracking of hosts or know from experience a host with really solid uptime?
I'm needing a new box for hosting a few static sites with Nginx (about 15 for now), and I really would favor uptime and reliability over the performance.
Also, what would your minimum RAM recommendation be for hosting this amount of sites? (Let's say 30 static sites max)
Comments
How many visitors? 128MB might work depending on that, if not 256MB.
Related uptime discussion, might be useful to read
http://www.lowendtalk.com/discussion/7036/your-vps-with-the-best-uptime
http://uptime.bluevm.com/
It's the system's uptime, not internet connectivity. You could have a box with 1000 days of uptime but yet a 50% connection uptime.
I think he means the connection uptime as 'uptime' ;-)
Yes. Uptime is also hardware uptime, both are solid.
the HW and network is normally 100%, it's up to your system admin skills to ensure the VPS up and running smoothly, as 128MB VPS will not run heavy loaded websites.
linode.com
racksrv.com
http://www.lowendtalk.com/discussion/7036/your-vps-with-the-best-uptime
Thanks, quite a good read.
% uptime over how long?
This - uptime as in a completely accessible machine.
Are you saying your services have 100% uptime? How do you guarantee that?
My sysadmin skills are more than capable of managing (if I even need to) some static sites (there isn't much)
I didn't specify that it had to be 128MB, asking for what recommendation are.
racksrv.com
Thanks (:
I like how you didn't even promote yourself
@Andre
My thinking is that if you are looking for great uptime you probably mean network AND hardware.
While I have some nodes with great uptime I have some which some may consider not so great, recently a UK node had a raid card failure now if I could have predicted that then I would not be a VPS Host I would be a millionaire stock trader with a crystal ball some time ago the Netherlands nodes had network issues which were completely beyond my control.
You will get hosts here saying "Pick mee pick mee" and showing the uptime of 1 node which is hardware up only but they are all subject to the same sods law, a disk can die at any time, an uplink can go down etc etc.
If you use racksrv or linode and probably a few others you at least have people that can physically lay their hands on the equipment, network or servers in no time at all and react more efficiently that others.
Counting great uptime and making a decision based on who has hardware running the longest is just a dice roll in my opinion if you want excellence you need to pay enough to allow people to provide it.
I don't immediately suggest myself because:
A) I don't offer DDOS protection
I dont have servers in high availability and self healing clouds
D) I dont have any physical access to the servers and I dont manage the network
At the same time I have nodes that have ran for a year with nothing but the odd momentary network 'blip' hence the dice roll analogy.
If you really really want 100% uptime you would need a fully DDOS protected high availability multi homed self healing cloud backed by some EMC SAN's with cross linked fibre, but that would be in the $500,000 bracket to set up. $7 p/month may not cut it.
If you just want as close as you can get while being prepared to live with murphy/sods law then yeah... Linode, Racksrv
my 2c
Almost every one of us can boast 98% or above. The real difference that matters to most people is how the outage occurs. A little network hiccup for 30 minutes at 3am your time isn't going to bother you as much as a 30 minute reboot at 7pm. I think there's a lot of relative factors that each person may find difficult to really communicate or take note of from anything other than experience.
With that said, I find Hostigation to be the most favorable for pure uptime of both network and hardware.
You make some good points. (:
Well to create the whole infrastructure, yes - but I think that a real 'cloud' solution would be the go, what do you think of DigitalOcean? Linode are a bit out of my budget at the moment.
Also, I might be needing 1 or 2 additional IPv4's for SSL, and DO don't currently sell additional ones apparently.
That's okay, but over what period of time? 98% over a month, or a year?
International visitors/different timezones - any downtime would bother me regardless (:
This isn't uptime-critical really, but it's most preferable within the budget. I'd be happy to sacrifice hardware for uptime.
Did they not recently have an outage?
They do have quite good pricing however and seem like a liable option.
I could do something on my new node once it's swapped for a supermicro
something
what
He did, but anyone would. How long it took to get there is a good question to ask, and it took quite a while for him to have that kind of outage. Even still, it was resolved within a reasonable time frame.
racksrv.com
Don't know about racksrv, but linode isn't as amazing as everyone says.
Only in LA.
No idea. Will start network monitoring when we've moved all locations to the new DCs.
Since I am with prometeus we had this:
1. A few DDoSes - in total I can estimate about 10 hours disruption (some packet loss in places, at times total outage) It only affects the target node, so, less than 30 min per server.
2. One IPv6 outage of some 7 hours due to he router giving up and at times there was no backup solution (IPv6 was not a priority till then).
3. One provider having issues with the line to us for some 12 hours. The link was intermittent causing routes to fail. Was only one out of some 4 including the back-up one (10 gb capacity but no commit) and except the routes to italy/middle-east and south asia failing from time to time and taking big detours for the rest of time, nothing was affected, services remained online.
4. Spurious reboots of OVZ nodes. Those were quite a few, Our announcements section gives an idea, but I would say, based on what i know plus Salvatore's tales, they were some 30, 40 at max for as long as 30 minutes since were only big nodes with a couple of exceptions (64-128 GB).
Biz services were unaffected by reboots/ddoses, but the ipv6 and route flapping issue did affect everyone.
Say, per customer, last year we had some 1-2 hours of downtime and some 1 day of various annoyances.
It is not stellar, but I would say pretty decent for our prices. Much over 99%
per month - yes it can be easily archived, per year it can be 99.999% (5 minutes of downtime) or less.
I used as example.
I dont have servers in high availability and self healing clouds
D) I dont have any physical access to the servers and I dont manage the network
I see what you did there...
Have personally found choopa one of the most reliable. Anyone there should be fine. Had a personal node there with 100% uptime for over a year no hiccups.
@Andre: Check the LowEndBox Top Providers list and you can't go wrong.
According to my experience, the @Maounique recap of last 12 months at Prometeus is accurate. The biz OVZ VPS has a better network, but the CPU of my biz VPS is older than the one of regular OVZ; and I experienced a couple of unwanted reboots on the the "low-end" OVZ VPSs last year, now solved. Support is fast and effective. In the end, uptime has been comparable to the OVH small dedicated server I previously used, and customer support has been better. I selected Prometeus also because they have good business references; they also serve relatively big customers outside the LEB space. My only concern is that Salvatore could get tired to offer hardware and services so good for a such low price.
I am not sure what you mean?