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
ethtool eth0
You can start check the physical port speed to start with, for example:
ethtool eth0
Then you'll need to find a local server that you can FTP/HTTP from and see what performance is like. You're better to find something locally otherwise your provider may say that connectivity to the site you are downloading from isn't optimal and thats causing the low performance.
Cachefly is often well connected though so running a simple download like:
wget -O /dev/null http://cachefly.cachefly.net/100mb.test
will give you some basic data.
iperf is another options, but be sure to factor in the overhead.
I tested on vps it says
Settings for eth0:
Link detected: yes
on VPS you can't see the sync.
Run a DL test and see speed on that.
Run more than one. We have clients come to us and they're like "WE RAN A SPEEDTEST TO A SERVER IN INSERT RANDOM LOCATION HERE AND WE ARE ONLY GETTING 12 Mbps!!!".
That's because you're testing against a node in the middle of no where and it's possible it's on the other end. Test a few locations and see. Just because you have a X Mbps port doesn't mean you're going to get X Mbps to every location. You won't always max out your port to everywhere.
Try iperf, just don't run it multithreaded on a vps or you'll be suspended.
Google how to use it.
Example: 2 digitalocean VM-s can push 1Gbit traffic between each other both ways.
Settings for eth0: Link detected: yes
As @Steven_F mentioned, try running multiple tests to get an average speed. It's very rare you'll get max connection even though the portspeed might be high. Run multiple tests, get the average speed and that's more or less what you can expect your connection to be most of the time
Try command with OS below:
CentOS: ethtool eth0 or mii-tool
FreeBSD: ifconfig -m
Guys, he is asking about the port speed of a dedicated server, not the network speed. Your bandwidth tests won't tell him the port speed, ethtool / mii-tool will tell him.
Actually he doesn't seem to know what he is asking. It sounded like he had a dedicated server, now it sounds like he has a VPS with a 20Mbps rate limit from a provider who has dedicated server with 1GE.
I interpreted it as him asking if he can tell if he'll actually get ~1 Gbps.
@MarkTurner you are right, what he really wanted to ask could be lost in translation.
That he answered in his opening statement, he said 20Mbps. So he already knows he wont get more than 20Mbps.
If you will only get 20 Mbps (as a guaranteed top speed) by contract, why should you care whether the general port speed is higher? You will never reach it anyway because of the restriction.
Or do you get 20 Mbps at 95% on a 1 Gbit port? Then your question would make sense.
Try installing an iperf on another VPS of yours. Preferably one with 1Gbps speed and geographically close to your VPS in question.
After installing run command iperf -s
Then, go to your VPS. Install iperf here to.
After installing run command iperf -c OTHERSERVERIPHERE
Try to see if you get past 20Mbps
I read it as he has 20 Mbps on a GigE.
I do not trust the publicly available speed test applications. I usually test the speed of my servers by SCP'ing a large file to the closest system available to that node. This is usually a good indicator of the full speed available.
And is inaccurate as scp requires much more resources to process the same size test file. Iperf would be the best option.