Speedtest CLI. Run speedtest on pfSense box
-
@johnpoz got it, it doesn't work for you. Did you notice that a widget has an uninstall (-) button?
-
@aln After some fiddling, I thing the problem is on how you interpret upload speed
It is completely wrong. It reports 26.6 when both trafic graphs AND running fast from browser reports upload around 4.
For test, I just inserted an$upload = number_format($info['upload_content_length'] * 8 / $info['total_time'] / 1024 / 1024, 2);
$upload = 5;effectively ovewriting the result, and got it reported after a speed test
Download speed and everything else seems to work
(and the time is completely wrong too, but thats minor)
-
@netblues thanks for the feedback!
I don't think that the calculation is wrong cause thats also how download is being calculated. I suspect that I push a too large upload that might be rejected by the server.
Can you check if the latest version works for you? -
@aln Well, its closer, at 6.6 but still not the real one.
Also it took ages to complete, and fails to keep the download steady too.
The cli version produced far more better results on the same network/hardware..
-
@netblues I think you are right, I'll scrap this version.
-
This is a dual core, Intel(R) Celeron(R) CPU N3050 @ 1.60GHz
Current: 1600 MHz, Max: 1601 MHz
2 CPUs: 1 package(s) x 2 core(s)
AES-NI CPU Crypto: Yes (active)speedtest-cli on pfsense and via browser is close too
1st peak is with cli, 2nd via browser
Hosted by Vodafone - Panafon S.A. (Athens) [0.97 km]: 36.415 ms
Testing download speed................................................................................
Download: 41.65 Mbit/s
Testing upload speed......................................................................................................
Upload: 4.08 Mbit/sand web based below
-
@aln said in Speedtest CLI. Run speedtest on pfSense box:
it doesn't work for you. Did you notice that a widget has an uninstall (-) button?
It works for me.. Your code is fine, an its great work.. My point is this not something that is needed or even wanted by anyone that actually runs a router.. And doesn't run a desktop sucking 95 watts sitting there idle as their "router"
Your point of testing it at the router makes more sense because it removes stuff that could slow down in your network is not a valid logic..
When you do a speed test, you should testing from your client to test your "network" not just pfsense.. Pfsense is meant to route/firewall traffic - not be a client..
I have made my point clear.. And here is what I am going to do.. Every single post that comes up that I come across asking about why is pfsense slow I am going to forward to you to answer ;) hehehe
@netblues your on dialup.. So yeah its going to be fine.. 40mbps?? That is an internet connection? Or your hot spotted off you phone on 3g?
-
@johnpoz This is a home lab which I use for tests, and no, its not dialup :P
Obviously, a dual core n3050 can handle 50/5 speedtests at ease.Getting back to the original speedtest@pf debate, consider sitting at the porch, and then the youngsters complain about games lagging.
Wifi is at limits, so instead of getting up and head to the home data center management room, one logs in to pf from phone. And since speed test from mobile wifi won't cut it, speedtest widget will tell you if you really need to get up.
Now I could run ssh terminal on android and use a bluetooth foldable keyboard and......
No.ps If speedtest widget is not available on a default pf install, and one has to find and install it by reading the forum, then all is good. I doubt anyone will ever complain about speed. At least not from the widget. I even bet a few beers on that.
-
@johnpoz your point of view is very narrow. There are different users out there with different needs.
Also if they came that far to use pfSense then probably not all of them are "fools". -
@aln agree totally, but at the time it gave me what I needed so thought I'd share it with the community. didn't think it would divide the community in the way it did. bottom line in my book is that if you are intelligent enough to find the solution to a problem you have you should be intelligent enough to fault find it. I'm not impressed with the negativity coming through some of the posts on this thread.
-
@tman222 it won't work i looked at it, it's compiled for a newer version of freebsd than that which pfsense runs on, unless of course it's been updated over the past few months.
-
@johnpoz all of my Google WiFi access points run daily checks. They are intended to be routers. I can't understand why you would say a router is not intended to do such a thing.
Have you considered that the line you see and speeds from the speedtest.net are not actual, untampered results? I see some pretty serious QoS problems on my provider.
Also, if you run your own speedtest.net server (like ISPs) there are ways to fake the results, like increasing file sizes (more mb = assumed more MBps for some reason) and setting priorities to known servers.
-
@adamoutler said in Speedtest CLI. Run speedtest on pfSense box:
Google WiFi access points run daily checks. They are intended to be routers
No they are not.. Do you mean the router, the nodes are not routers...
They have their own built in speed test, and just had my son run it on his - guess what, it doesn't show his full speed 500Mbps he is suppose to get from his isp..
Is that isp, is that the google router test... Not sure - since he is just moving into the house.. And he has nothing that can actually test via wire gig.. So that will have to wait.
Those tests you posted or horrible!! 346 one day and then 3.5 the next? What speed do you actually pay for?
speeds from the speedtest.net are not actual
WTF does have to do if run from router or client behind router... Point has been made - you guys have fun... If you want to run speedtest from where ever... have at it... Point has been made.. I just tied of seeing why is pfsense speedtest not showing full speed... Which as been asked for YEARS... And this is only going to bring more of those..
You know how I know I am seeing full speed of what I pay for, I agree with you - speedtest nests can be misleading.. Is when I download files from my server in the NL.. .And see 60+ MBps which is my 500Mbps I pay for from isp..
-
@johnpoz point the finger where it belongs. Point it at the ISP. They aren't giving clients the actual results. They run their own servers and you should trust pFSense over an Speedtest.net which is paid for by ISPs.
-
Dude has Zero to do with the ISP... What do you not get that a router with X cpu designed to route packets might not be able to do 500/1000 as a client?
Have already show this multiiple times... Test to same freaking place ran on client shows full speed, test on pfsense does not will..
I'm done - have fun...
should trust pFSense over an Speedtest.net which is paid for by ISPs.
You clearly have no clue to what is even being discussed here to be honest..
-
@johnpoz if it can't do 500mbps as a client, then how can it possibly act as a client AND a server for packet filtering, suricata, or plain routing? This doesn't make sense unless something was seriously wrong with the libraries.
-
Hi know this topic is old but I did not see any other updates elsewhere.
@aln Thank you so much for the work that you have done with this! While there is allot of blustering from people on here with their personal options I think this is a handy tool.
pfsence has many more features then just "routing" and the more information you have about network speeds from different points can be useful.
Personally I like this tool because I have a pfsence box that can handle the workload without issue. It is useful to me to see what speed my edge device is getting to the outside world. (without the fuss of going though the other hops in my internal network) It also helps me keep track of my crappy ISP.
That being said If you ever wanted to make an update.. I would love to see something that gave a history of over period of time that would allow you to see to trends etc.
Just an idea, not sure if anyone else would like this etc. But once again thank you!
EDIT : kind of like what @provels has done but without using a 3rd party etc.
-
@Exilewolf
It was a fun exercise, but the results are so variable that I don't think the data is really worth the bits it's printed on. The big dip to the right was easily recognized without the data. It was a local outage. FWIW, this data was gathered from a server behind the FW and not just the FW. I have a 300/20 line.
-
This post is deleted! -
@aln thanks so much for this. can you please clarify how can i bind an specific interface when running a test from your widget?
