What is standard industry for server uptime?

What do you expect from a decent host of your server (in our case a dedicated server)? 99% uptime? 99.9% uptime? Don’t mean to sound petty with the percentage but recently our server crashes at least once a day and it takes anything between 10 minutes to 3 hours for our support to get it back online. I was wondering if it’s too much to expect no daily downtime?

Daily crashes are unexcuseable. In today’s market you can get whatever you want.

I know a few people (who have dedicated boxes) who have 100% uptime for 6 months or more.

Then their servers are highly insecure then if they have had 100% uptime for 6 months on a single server, as certain patches require reboots, be it a Windows set of patches or a Linux kernel upgrade. 100% uptime is doable, but only with a load balanced cluster of servers, so that there is always one available when the others are being patched. As for daily downtime, there’s not much of an excuse for that really.

Daily crashes are a bad sign.

They say it’s due to server overload because of my VB forums. My traffic has doubled in recent months, so maybe I need to upgrade.

Are you buy any chance using mod_gzip?

I really don’t know - what is it?

In that case, most likely not. It is a module for Apache that gzip compresses all textual output to a visitors browser, providing it supports it. It can often lead to crashes though if you’re not careful, due to it filling up the /tmp partition.

Good thinking, I would suggest you go with a decent host that will allow you to use a bit more resources at a little extra cost or a dedicated server.

Example: We buy our dedicated servers (Example Server name: Server 2) before we completely fill up the current server (Server 1). So if we know a bigger site is moving over onto our servers we will put the big site on Server 2 instead of server 1. This evens out the balance and avoids overloading servers.

They are already on a dedicated server. Upgrade wise, I’d check you’ve got enough RAM, then I’d think about adding a 2nd server to run your databases from.

A 2nd server! You think that’s needed already? Interesting question - I think I’ll start a seperate thread about it.

Have you tried optimizing your web server (or even database) settings to see if it fixes the crashes? I’ve seen some poorly configured servers in the past that have really affected the performance and stability of the server.

I’m learning a lot during our search for a new host. Apparently our sites aren’t considered that big (need the shy smilie here :slight_smile: ), and many have offered us practically the same hardware that we have now. So, I guess our server isn’t optimized at all.

One of our servers has had 100% uptime since it was put in with the exception of two reboots because of kernel upgrades (which only took about 2 minutes each).

That server has been in production for just about a year now. Others have 99.8% uptime and up.

I think that any host can have a problem with a machine… leading to a period of downtime. But it shouldn’t be repeated. They should solve the problem and get the server back up and running as quickly as possible…

As someone said above… daily crashes are just unacceptable.

Well, usually when hosts put in uptime charts, they use network uptime. Maintenances are a different story.

Most hosts offer, 99.9% uptime, some 99.99% uptime, and some hosts just offer 100% uptime with a SLA.