How to troubleshoot intermittent, slow speeds over the internet?
Linux - NetworkingThis forum is for any issue related to networks or networking.
Routing, network cards, OSI, etc. Anything is fair game.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
How to troubleshoot intermittent, slow speeds over the internet?
Server A is Debian 12 and is located at my house directly plugged into an ASUS AX86U router.
Server B is Debian 12 and is located at my brother's directly plugged into an ASUS GT6 router.
I have AT&T 1 gig symmetrical and he has Comcast 1.2 down/35 up.
I recently moved my backup server to my brothers house and -sometimes- when the backup runs, we get about 45-50MB/s but MOST of the time, it's about 3MB/s. In a perfect world, we should be able to get about 90-95MB/s (maxing out my upload speed) but the internet being the internet and peering agreements sometimes being garbage, I was hoping that the 45-50MB/s would be consistent but it's not.
Even before I moved the server to his place, FTP and iperf testing would (on a great day) max out around 50MB/s but most of the time, 3MB/s.... 5MB/s..... just crummy speeds.
Since we're both residential customers, I know that trying to get AT&T or Comcast to solve this problem would be pointless but what testing can I do to 100% rule out anything within our control is causing these totally random speed fluctuations?
I believe the theoretical max speed on a 1GB ethernet is something like 943 Mb/s, 117 MB/s on a typical home network.
So 35 up is probably Mb/s and so at times I might expect peaks at 45-50 Mb/s but that is still an average of about 4 MB/s. As far as I know rsync should use the available bandwidth. Is it because rsync is only as fast as the slowest link?
I believe the theoretical max speed on a 1GB ethernet is something like 943 Mb/s, 117 MB/s on a typical home network.
So 35 up is probably Mb/s and so at times I might expect peaks at 45-50 Mb/s but that is still an average of about 4 MB/s. As far as I know rsync should use the available bandwidth. Is it because rsync is only as fast as the slowest link?
It's likely because your next hop router is shared with many.
What technology is provided by your ISP for connectivity does matter.
Is it xDSL, FTTH, Wireless, etc?
It's likely because your next hop router is shared with many.
What technology is provided by your ISP for connectivity does matter.
Is it xDSL, FTTH, Wireless, etc?
There's also traceroute. You can check where you're going. Long ping times = heavy loading.
I once subscribed to a site which was being served by an AWS server in Dublin's Citywest Industrial Estate, not 5 miles from me. But my route to it went through trunk servers to the Excited States, through some bottleneck, and back from the Excited States to City West
Mind you, I did find that one of it's 3 web servers gave me a more sane route, and used that for the duration of my subscription.
There's also traceroute. You can check where you're going. Long ping times = heavy loading.
I once subscribed to a site which was being served by an AWS server in Dublin's Citywest Industrial Estate, not 5 miles from me. But my route to it went through trunk servers to the Excited States, through some bottleneck, and back from the Excited States to City West
Mind you, I did find that one of it's 3 web servers gave me a more sane route, and used that for the duration of my subscription.
Ha, yep, sometimes routing is crazy.
A traceroute to my brother didn't reveal anything horrible. The max time I saw on any hop was 30ms.
I've more than once used it to prove the paid for service was bad or the router was crap.
It's actually fun to use the combinations of packet sizes, intermittence and tcp or udp.
Then printscreen and timestamp your evidence.
I've more than once used it to prove the paid for service was bad or the router was crap.
It's actually fun to use the combinations of packet sizes, intermittence and tcp or udp.
Then printscreen and timestamp your evidence.
All good.
Try changing the packet size to default.
And also change the interval to default.
For more info try in cli: man mtr
Is this the culprit? IP: 32.130.89.25 (see attached image... don't know how to paste it 'in-line' as part of the message.)
No idea what it is... looks like it belongs to AT&T. How would I go about getting them engaged or to even care? Or maybe I don't know how to interpret the data and what I'm seeing is normal?!
No idea about AT&T and ISPs in the US but would not be worse than here in Aus I hope.
What I like to do when performing these tests is to transfer a huge file at the same time.
Take for example the FreeBSD amd64 iso, it's over 4G. https://www.freebsd.org/where/
Then perform tests with mtr while transferring direct from the hdd of a local laptop to another hdd on the remote server.
The tests are done with tcp and also udp.
Don't forget the time stamping for every test.
Once the evidence is gathered, then escalate to your ISP(s) Customer Service teams.
They should have SLAs in place for paying business customers.
I'm curious: Are you and your brother in separate towns?
No idea about AT&T and ISPs in the US but would not be worse than here in Aus I hope.
What I like to do when performing these tests is to transfer a huge file at the same time.
Take for example the FreeBSD amd64 iso, it's over 4G. https://www.freebsd.org/where/
Then perform tests with mtr while transferring direct from the hdd of a local laptop to another hdd on the remote server.
The tests are done with tcp and also udp.
Don't forget the time stamping for every test.
Once the evidence is gathered, then escalate to your ISP(s) Customer Service teams.
They should have SLAs in place for paying business customers.
I'm curious: Are you and your brother in separate towns?
We live in the same state and are about 2 hours away from each other.
Since we're both just average home users, it will be difficult/almost impossible to get an issue like this brought to the attention of somebody at my ISP with the power to fix it. From my limited knowledge, it appears that whatever that hop is, there is massive packet loss happening there.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.