There is no SQM. This issue is totally out of anything I can imagine what could cause it, because like I said, it doesnt happen with all downloads, but just with a few, and I isolated examples of where it works, and where it causes this effect. And it is even weird, I try to explain:
cable router (500mbit / 64mb/s) <-to wan> Openwrt router <- to lan> lan
starting a wget -O /dev/null http://speedtest.tele2.net/1GB.zip from any LAN side I always get max speed, regardless how often I start the wget, how fast I quit it with ctrl+c no matter what. ALWAYS works
starting a wget -O /dev/null http://v4.speedtest.belwue.net/1G from any lan side, the above issue is triggered, download may start or may not with max speed, then after 2-3 seconds is degenerating to around 5-10mb/s and then repeating it it mostly also starts with just 5mb/s
starting the same wget -O /dev/null http://v4.speedtest.belwue.net/1G DIRECTLY behind cable router, it works fine, always, always max speed, same es example 1.)
This doesnt make any sense to me. It is not a always broken issue behind openwrt, but just with SOME servers, which though then work FINE, when used directly behind cable router. How could that even be possible?
Isnt all download or data the "same", just data, just 0 and 1, how could one lead to this issue, and one not? I have tested several download test servers, and some trigger it, some not. But they ALL work correctly directly behind cable router, and some are broken in the line cable router <-> Openwrt <-> lan.
For about 30 minutes or randomly 20 tries it suddenly worked after 5 reboots of the OpenWRT router and now it is again broken. This seems to be some weird nasty bug. Maybe double nat has something to do with it?
This cant be a SQM issue. I tried giving one of the Windows PCs the IP of the Linux NAS which has no issues, and it triggers the same result. It has to be some bug with the OpenWRT Linux or/and in combination with Windows 10 Clients. My theory is, that is somehow has to do with the TCP receive window and the tcp autotuning algorithm of Windows. I am not an expert in this, so it would be welcome, if someone who is, could say something about it. I looked into it the past day and as I understand, there is no RWIN anymore you can set manually in Windows 10 and it just uses an autotune algorithm for the tcp receive window.
If I actually do a "netsh int tcp set global autotuninglevel=disabled" on the Windows clients, my download rate instantly become stuck at perfectly same rate of 3MB/s and never goes over it. This is not normal too I think. And going back to the default value of "netsh int tcp set global autotuninglevel=normal" results in the issue above.
I read into the tcp autotuning algorithm and it seems to work like this, that it communicates with routers and the end server, so maybe here is the issue somehow, that OpenWRT or the double NAT somehow blocks this autotuning "talk" but I have no idea how. Or that the double NAT somehow interferences with it, that it adds some latency to the timestamp calculation of the autotuning calculation, and then Windows bumps down the TCP receive window on the fly and the above issue happens. This is just a theory of course I am not an expert in this. Anyone maybe can say something about it?
MS says about that: "When the receive window autotuning feature is enabled, older routers, older firewalls, and older operating systems that are incompatible with the receive window autotuning feature may sometimes cause slow data transfer or a loss of connectivity between Vista clients. When this occurs, users may experience slow performance. Or, the applications may crash. These older devices do not comply with the RFC 1323 standard. Some device manufacturers provide software that works around the hardware limitations."
You asking about CPU load doesnt make sense if you read what I said 100% through and look on all gifs carefully.
Just a quick question. I am getting ready to upgrade from kernel 4.14.103 r9506 to 5.4.41 r13342 of David's builds. Any gotcha's I need to know about (1900ACS)?
(I was advised download a backup of my config, remove /etc/config/ubootenv and /etc/fw_env, repack. Upgrade w/o keeping config, then restore the modified backup.)
I did a quick search on the autotuning side of things for my own benefit, as I'd not heard about it before.
An app was noted in one of the searches "SG TCP Optimizer". Whilst I wouldn't normally want an app making changes, I saw this at least shows the PowerShell etc commands before it commits.
I'm only throwing this in as, to me at least, it shows a lot of options for configuring the stack which may be worth you looking at?
Obviously I already tried the TCP Optimizer, with no effect, sometimes even seems as it making it worse. Here is another gif showing the issue, this time a periodic up and down on Windows 10 client, where on Linux it is 100% stable 64MB/s:
I wonder if this is a bug in Windows 10, which maybe wasnt there long ago, and no one noticed so far, or a bug in the OpenWRT image, maybe of the new kernel.
Hi all,
I'm just about to setup an OBI302 ATA with freephoneline on my OpenWRT router. According to this guide http://forum.fongo.com/viewtopic.php?f=15&t=18805#p73839, I have to check the UDP unreplied timeout and UDP assured timeout for values lower than 20 and 120 repectively. Does anyones know if I have to change any of theses settings and if so, where to set theses. What are the defaults values in Davidc OpenWRT, I see somes files in /proc/sys/net/netfilter but there all 0 byte
Thank you
I've found file 11-nf-conntrack.conf in /etc/sysctl.d with theses values that can be modified in /etc/sysctl.conf:
net.netfilter.nf_conntrack_udp_timeout=60
net.netfilter.nf_conntrack_udp_timeout_stream=180
Just a quick question. I am getting ready to upgrade from kernel 4.14.103 r9506 to 5.4.41 r13342 of David's builds. Any gotcha's I need to know about (1900ACS)?
(I was advised download a backup of my config, remove /etc/config/ubootenv and /etc/fw_env, repack. Upgrade w/o keeping config, then restore the modified backup.)
Hi! I have installed OpenWRT 19.07.4 for my WRT3200ACM from official site and now I'm looking to have luci-theme-atmaterial but I can't find the IPK for Luci, I have Luci git-20.272.48698-9883bb3-1
Guys, only 85 people signed the petition in nineteen days. During this time, this most popular thread dedicated to WRTXX00, probably, was visited by more than one hundred owners of these devices. Why are we so passive, everything suits us!?
Let's be active!
I am simply trying to restrict one device on my network from accessing the internet. I have done this many times before, but it doesn't seem to be working this time and I think it is best if I get another set of eyes.
The destination address field does not have 'any' as an option like it used to. I am wondering if this is the problem, and if so, what is the numeric equivalent.