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.
uptime < 5 days
Hi, I've seen threads here about comparing server uptime, but I don't get ALL the reasons why a small uptime is not a good thing. That is, I have a VPS with WeLoveServers, and uptime never exceeds 7 days (i.e. node is probably rebooted every often, varying from 1 day to 5 days).
Apart from being unavailable while rebooting, what are the other bad things about a VPS rebooted so often at the node level?
Comments
Imagine you are running a DNS server from your VPS, do those few minutes its down still seem like nothing?
DNS server is one of the web services I run from this VPS. I don't know exactly how DNS resolution works when one of the DNS server is unavailable (does the client wait for some minutes before deciding to try the alternate DNS?), but unavailability on reboot seems to be a non-issue on DNS or mail service so far.
You must have your DNS with failover as you should then. Basically when a reboot is in progress anything on the box is down, no emails get routed (unless failover setup) same for DNS and web.
Now depending how mission critical your application is determines if this is acceptable.
Its still a shit state that people are asking if downtime and reboots matter.
I would hate to be disconnected from IRC networks I use every few days. That's just annoying. Some oldest IRC networks don't have chanserv to keep channel ownership and people use IRC bouncers to stay connected not to re-connect every few days.
It really depends on what you do with the "uptime", if it's production and no failover/redundancy then downtime would mean impact on productivity which will cost.
On the internet a lot of people won't wait 10 seconds for a reboot. They'll see your site is down and seek out a competitor.
A restart in and of itself isn't bad burt shouldn't be necessary every week on a well maintained system.
All I have to say is most of my server are currently have great uptime, like 200+
https://www.munroenet.com/uptime/
Check them out above.
You should add an averages for the stats on the bottom
hmm, don't really see the need other then e-penis, should that really be a feature?
OP, do you really not understand the importance of having a good uptime?
What's the point of paying for a server hosted at a datacenter facility if you can't have it available 24/7/365? In this industry, uptime is of utmost importance, and wether or not you use your server for something that is mission-critical, it is not support to be down so very often.
If your hosting provider reboots a node every week, this is a bad sign. Either their hardware is inadequate or they oversell and overload the bejesus out of it. Either way, it's bad.
IMHO, the "normal" uptime is between 99.999% and 100%. You don't even need to reboot for OS updates and patches anymore when you have tools like KSplice and KernelCare.
If your uptime does no exceed 7 days in the row, then this is just plain ridiculousness.
If you have a redundant DNS Cluster, DNS Resolution may take a few milliseconds to a few seconds extra, which will obviously effect a page's load time.
If you're hosting email, the message will bounce back, obviously. And depending on whether or not the server you're sending from supports retries, the message may or may not be delivered when the server is back up.
If you're fine with having a downtime every few days, you might as well get a APC UPS device, a few static IPs from your internet provider, and host your servers at home. And even if you do this, I guarantee you, you will have better uptime than 7 days.
IMHO, your hosting provider is not a good one. But if you like it and you're fine with the frequent downtimes.... Oh well.
Have you raised a ticket with WeLoveServers?
You should be concerned about uptime. Most of my VPSes have 50+ days of uptime. One currently has 8 days of uptime, after 44 days of being up. The node needed to be rebooted due to a kernel issue and an email was sent out before the node was rebooted.
Depending on what the provider's support says, you might want to switch to a different provider. I've been with RamNode for a while and I haven't had any issues with them. Uptime is great, support is great, and the prices are reasonable. They might not be dirt cheap, but it's worth paying a little bit more for.
how do you use most of the space on your boxes? do you backup others onto them, or do they just have little disk space?
That's free space.
Ah, that makes sense.
KSplice and KernelCare again and again prove that they are not good replacements for actual kernel maintenance updates and emergency reboots, the recent OpenVZ exploits are good examples of this. It took KernelCare a good 6 hours after public release of the exploit before a patch was available, in which case; people would have already leaked credit card data and other related material; had hosts waited for the update.
Thank you all for your reply.
1] I raised a ticket to WLS weeks ago, asked them with the same dumb-like tone as this post and I think I was replied accordingly. That is: the answer suggested frequent reboot is not really an issue.
2] The country I live in would only give 95% uptime for a server hosted at home and speed as low as 1Mbits/s.
3] I host on the WLS server few production web sites + email, and a backup of other sites.
4] Server purchased on a black friday sale, I now understand the reason is overselling.
@CharlesA: from a provider, sending a mail before a node reboot is a proof of quality, I'd like to send mail to my end customers as well to prove them the 1mn downtime is a maintenance necessity.
Conclusion: I will switch provider. Actually I'm a satisfied Hostigation & Ramnode customer, but their price per hard disk space is maybe not the most low-end, and my server mostly needs hard disk more than Ram and cpu.
@DotMG: How much disk space do you think you'll need? I've been running a 128MB box hosting a couple services with 80GB of space. It's OVZ, though, and I know KVM has less disk space allocated than OVZ.
5 days! We have a hypervisor node in London with the following
09:53:55 up 951 days, 17:00, 0 users, load average: 0.33, 0.15, 0.11
Customers on the same node as my server would be happy with this kind of usage I do:
Not that bad tbh. 80GB drive on a 3GB box. It doesn't even look like you are using anywhere near the 3GB either.