Scanning library takes hours just for a couple of movies/shows

The more I think about how routers are configured,

If they perform L2 switching, they use MAC addresses.
For this application, they need L3 switching, which are IP addresses.

They do need L2 switching at some point because your modems are managed by them at the MAC level.

If someone just fat-fingered a config somewhere and pushed it out to several routers, this would happen.

Certain MAC address ranges (tags) would work, others would not.

If all switches were properly IP (L3 switching), this would never have occurred.

I understand the DSL problem very well.

DSL = 2-10 Mbps
ADSL = 6-12
VDSL = 25
VDSL+ = 50 - whatever (depending on length)

All of this subject to pair bonding. G-bonding is the most common.
It’s what I have here. 2 pair, G.bond, VDSL 20 Mbps each (due to wire length), yielding 40 Mbps.

Just to make sure.
We are not talking about DSL it is Cable Internet. (Coax 1GBt).

I understand. I’m also sharing that I do understand the DSL side of the equation too (slower speeds and matching taking longer) due to the gross speed differences.

I had to buy two VDSL lines to get 50 raw (40 payload) Mbps of service.

I don’t think it is deliberate. I just encountered another problem. The login screen in the Geforce Experience Application (utility application for Nvidia GPUs) could not be loaded. Like for the media server I was able to fix this by using a VPN.

Then it’s a fat-fingered config somewhere on a rack. Good to know

Be certain to mention that to them…

Anyone a good idea How to properly report that? Is traceroute the best tool/way ?

Can we know where in that traceroutes exactly the wrong turns are?

If we have I will post our results to the Vodafone Service forum in addition to @reap s technician call (which will hopefully happens soon )

Reap has the best numbers.

He used his Mac and his Syno (both on the same LAN) and got different results.

That’s not possible with properly configured routers & switches at the ISP

Sharing the WAN IP, LAN IPs, and MAC addresses will give them info to go on.
Even the MAC address of your modem is probably helpful. They must have some form of MAC/VLAN configured somewhere

In a german thread about this topic someone mentioned that he has no problems with the Vodafone station unlike the Fritzbox. Can anyone reproduce that? Unfortunately I only have the Fritzbox.

I am a GigaMax customer in Munich with my “old” Fritzbox 7590 and a Vodafone station in bridge mode.

I don’t have any of the problems with loading posters or other meta data.

1 Like

If I wouldn’t need my box for homeoffice I would suggest to bring my box to test it. This is absolutly weired behaviour.

And still I have no valid testcase which should work and is failing constantly to hand this out to vodafone. :frowning:

Sorry guys, took me way too long to answer. But SARS-Cov-2 got me working way too long the past couple of days (working in healthcare).

So a guy that called himself a technician called me on saturday morning. What can I say? He told me to gibe my 2,4 GHz and my 5 GHz network different names and change the 5 GHz from Auto-Channel to 36-50 MHz. No bull****, that’s all he said. I told him everything @ChuckPa and I discovered and that there are multiple people with the exact same issue. Even then fact that I can prove everything I say with screenshots or logs of the results didn’t convince him at all. You just can’t imagine this…
Believe it or not today I really tried this bs. Guess what? Didn’t change anything…surprisingly.

At this point I really can’t tell what to do. @bwestpha wrote me a PM with the idea of posting in the forum of the ISP which I think is our best bet right now.

@Coxeroni 6590 means you’re still using DOCSIS 3.0 instead of 3.1 which could be the whole point of this problem.

7, not 6. 7590, a non-cable FB I use after the cable modem in bridge mode. Needless to say that the modem operates in DOCSIS 3.1 mode. At least I guess so, the modem itself does not show a lot of information on its web interface.

Not for me to stick my nose in too far but:

  1. demand escalation because WiFi setting has ZERO impact on routing a WIRED connection. IDIOT
  2. What’s your legal recourse should they fail to provide contracted service?

I believe that this problem has something to do with newer Cable Fritzboxes (6591 and 6660).
Hopefully it will be adressed with a firmware update by AVM.

@reap @AdmiralAckbar73 @ChuckPa @preacha @xLuchs

I solved it. Guys… really. It’s working again.

This is a combination of multiple factors. If:

  • you are using AVM FritzBox 6591 or 6660 and
  • using any new Vodafone subscription which involves DOCSIS 3.1 and
  • have a FritzOs which is quite new (>7.10 or so)

this problem hits you.
The solution is quite “easy” if you know what happens and I stumbled uppon this solution by accident.

I tried to capture the network stream between my host machine and the internet via FritzBox and usually there is a button for that which outputs you the . cap File (anyone remember?).

Since a few FritzOS upgrades this buttons dissapeared to an unknown menu :smiley: but nevertheless I found it :

Guess what? It won’t work with my AVM 6660 - its too new and the buttons all do not work to capture any traffic.

But the journey was not useless… please guys. Scroll down a little bit in the last screen:

It’s called “Paketbeschleunigung” (Packet Acceleration). Disable “Hardware Beschleunigung” - and you are done. This is it.

I think there is something missing in the AVM implementation of DOCSIS 3.1 in combination with that hardware acceleration.

I have not found more information about that acceleration but a topic here: Was genau macht Paket-Beschleunigung und Hardwarebeschleunigung? | IP Phone Forum and some sentences remind me of what @ChuckPa told me

…da werden dann IP-Adressen und auch -Ports umgeschrieben bei portbasierten Protokollen und ebenso die “L2-Adressen” (aka MAC-Adressen), die ebenfalls passen müssen, damit das richtige Ziel im LAN dieses Paket überhaupt erst erhält…

Wait what? L2 adresses, MAC filtering bla bla?
→ directly disabled the acceleration and it worked for me.

… too much magic ain’t good …

Man, what a pain in the ass !

7 Likes

Great find.

Perhaps you should forward to the “Change the WiFi” guy?

:rofl:

I’m going to test this as soon as I have time to.
If it really works this should probably be forwarded to AVM.

I hope this is the solution for most of you guys ! Let me know.

@bwestpha OMG you are a hero. It workes for me, too. Posters and metadata are loaded immediately again. Also the pictures of the cast that was missing before.

Thank you so much. Please let us know the feedback from AVM.

@preacha great. This is nevertheless a temporary solution since - already someone else reported - this Leads to slower transfer speeds.

Im still courious what exactly Plex (to be more specific, the python2 request module) is doing differently in comparison to an simple wget call from the same machine.

Maybe tomorrow I will try a sample python program which should try to reproduce this behavior. Maybe in addition with wireshark if I get the capturing working on my 6660.