HI to all i hope to find all of you well.
I have a problem with nethserver if I test the line with a laptop it marks me 85mb download and 65mb upload
if instead I do the test passing through the firewall it marks me in download 85mb and in upload, sometimes 3-4mb other 7mb the maximum that arrived was about 22mb.
I tried to disable the various services one by one but I found no benefit … any ideas?
thx a lot
Hi, welcome to NethServer community.
Which speedtest did you use? The result may not be accurate.
Maybe there’s a wrong upload in WAN settings?
Which modules/services did you install?
hi i use speedtest.net and the intenal speedtest of nethserver (speedtest-cli)
the services i enable are:
firewall
ips
jail
thread defence (blacklist)
web proxi
openvpn
i try to stopped one service at a time and took the speed test
hi i use speedtest.net and the intenal speedtest of nethserver (speedtest-cli)
the services i enable are:
firewall
ips
jail
thread defence (blacklist)
web proxi
openvpn
i try to stopped one service at a time and took the speed test
Maybe you enabled traffic shaping?
Did you check the WAN settings?
It would be interesting if a fresh installed NethServer without the services has similar upload problems.
Could it be a hardware issue? (faulty network card or cable)
i try with another cable but in these days i want to install another nethserver and make test.
is the test is good
i try to use another nic on the original nethserver as wan or reconfigure all services on the new one…
btw i can disable or uninstall the traffic shaping with no problem?
thx a lot
ivan
It’s part of the firewall module but you can disable it.
Where is the Speedtest feature available?. The Speedtest feature seems to have been removed or moved.
Also, the Applications > Firewall > Settings > WAN page now has the ability to pick Active-backup or balance. In Active-backup mode, the display of upload/download bandwidth seems less useful than the time graph that appears under charts when balance mode is enabled. But charts seems to have reintroduced an old minor bug, where speed is labeled KB, which is the convention for BYTES per second instead of Kb for bits per second.
Hi @harry
An old confusion when to use bits, when to use bytes.
→ Bits is used in transfer, eg over the wire, wireless, etc.
→ Bytes is used for storage, eg Diskspace, USB, and so on.
It’s simple, but there is still confusion about it!
As such, a Speedtest MUST use a bits denominated measure, be it kb/s, mb/s gb/s or whatever.
A speedtest for the USB-C speed would also be given in Bits (Gigabits nowadays), whereas the storage capacity would be given in Bytes (Gigabytes nowadays).
Then again, there also programmers using (suffering from) an autocorrect capable system…
Wrote kilobits, system changed it to kilobytes, but no one actually “realized” the error. Your eyes correct such errors, without registering them… 
My 2 cents
Andy
It has been removed, see The automatic calculation of the bandwidth with speedtest-cli lacks precision - #19 by giacomo
@Andy_Wismer I was not confused, but always happy to have people add value to a discussion Her is a tiny bug. The Firewall > Traffic Shaping, (sorry about incorrectly saying it was on the WAN page) the display uses a B as in Bytes instead of a b as in bits.
It seems correct to me. In that charts data are in Bytes and not in bits: ui. fix unit of measurement in bandwidth chart by andre8244 · Pull Request #181 · NethServer/nethserver-firewall-base · GitHub
It is still wrong.
Rule of thumb:
Transfer in Bits. Bits is in small caps!
Storage in Bytes. Bytes is capitalized!
This is almost universal - except for those who don’t quite understand universal conventions or the difference between bits and bytes…
As the joke went, when the subject was (nuclear) fusion:
There’s a lot of confission about fusion…
My 2 cents
Andy
A correct example from Proxmox:
Note the capital G in the memory (Storage), wheras small k in Network (Transfer).
I can agree with you, but you don’t know how many times we discussed to change it because users do not understand the chart.
Actual implementation is the one which currently generates fewest support tickets 
And…
Are the tickets displayed in Caps or small caps?
(Amount of generated support tickets…)
Should be in K, as objects in a stack are generally storage…
My 2 cents
Andy
Giacomo, I concede this is a minor point, but with respect to your claim it looks correct to you, see the peak transfer rate on my chart. It shows about 292MB. That would mean that my internet connection would be operating at about 3 Gigabits per second download, and . That is not possible.Further, the upload speed of about 18 Gigabytes per second is 10 times what I measure from my desktop.
A bit more detail: I tried I have actually seen throughput through my little fanless i7 CPU box of 1Gbps using Speedtest from my desktop and from the former Speedtest feature in Nethserver. People had expressed caution about the perfomance of the built-in RealTek NICs, and the i7 CPU, but when I ran those tests, the CPU usage was less than half on one core.
I had upgraded my Comcast service from 600 Mbps to 1Gigabit because most of the time, speeds were below 1Gigabit per second. Comcast did a mass improvement for this class of service from 600 Megabits per second download, and I consistently get 18 Megabits per second upload, versus Comcast’s advertised 15 Megabits per second best efforts for this class of service.
My bad, I didn’t understand the value was wrong, I thought it was only a about displaying the values in a different format.
This is my speedtest:
Retrieving speedtest.net configuration...
Testing from Telecom Italia Business (80.17.99.73)...
Retrieving speedtest.net server list...
Selecting best server based on ping...
Hosted by IRIDEOS SpA (Rome) [219.69 km]: 16.062 ms
Testing download speed................................................................................
Download: 79.35 Mbit/s
Testing upload speed......................................................................................................
Upload: 80.66 Mbit/s
And the relevant graph:
Indeed the graph is wrong.
Edit.
There was a correct issue reported by @stephdl: Firewall-ui: Trafficshapping do not use the good bandwidth values · Issue #6479 · NethServer/dev · GitHub
But we fixed without changing the label 
Firewall: wrong unit of measurement in traffic shaping bandwidth chart · Issue #6475 · NethServer/dev · GitHub
Edit 2
I’ve created a new issue: Firewall: wrong label in traffic shaping bandwidth chart · Issue #6623 · NethServer/dev · GitHub
Thanks @harry and @Andy_Wismer !
The package is ready for testing.
To try it:
yum --enablerepo=nethserver-testing update nethserver-firewall-base-ui
The package will be automatically updated to next stable release.
@Andy_Wismer @harry would you give it a try?
Installed. No issues.
So far looks ok…
GUI in Cockpit looks also good (so far…)
It may be more of a rhetorical question, but Disk-IO is actually transfer, not storage, so here (most likely) small caps would be correct, but I’ll admit, here I’m really not sure…
My 2 cents
Andy
