Slow connection time #54
Comments
|
It usually takes a lot less than that. Let's exclude a couple of things:
Also, the container logs is really welcome... |
|
Here is the log: After pressing the "Connect" button it took 17 seconds at the first try and 25 on the second try, until the "New relay request" message appeared in the log. If it timeouts no log entry will be created.
I tried to use IP instead of the hostname, but it doesn't make any difference.
Yes, I can confirm that the same key was entered on both clients. |
|
I assume the rustdesk server is on a VPS somewhere and the two clients are on different networks. Is it possible to test if these delays occours with this server and two clients on the same network? You can also check the timings for direct IP connection. Another idea is to use a (temporary) rustdesk-server inside the network, to exclude any router/firewall issue. |
That's correct.
If they are on the same network it takes 2-7 seconds. But there are no log entries, when they connect. Probably because the relay server is not involved?
I enabled direct IP connection and opened port 21118/tcp on the client. It takes 2-5 seconds to connect. |
This is our baseline, we can't expect anything faster than that. In this case, no server is involved.
So the previous case plus the time for hole-punching.
The log entries are from the relay server (hbbr) which you don't use because both clients are on the same network. In this scenario, only hbbs is used. To recap:
At this point we can exclude the clients and the internal network. |
I can connect to my PC in office from home with less than 1 second, public server (AWS, before it is vultr, Thanks for AWS's free credit) in Singapore. |
We should have a talk about connection latency and speeds around the world. From my office I can ping my external rustdesk server in 16ms (avg), I have 100mbit (down) / 30 mbit (up) FWA. This here (northern Italy) is a "good connection" (except with heavy rains, which causes disconnections). In Italy, FTTC is a hit/miss (mostly miss because of the long copper connection), FTTH is mostly unavailable outside big cities. |
When I do that, then I get the message I tried it with both docker-compose configurations I posted at the beginning, except I added
Connection CLIENT1 and CLIENT2 (same network): 250Mbit down / 40 Mbit up, 15ms ping to external rustdesk server |
|
Just to test all the cases, can you try this configuration: version: '3'
services:
rustdesk-server:
container_name: rustdesk-server
ports:
- 21115:21115
- 21116:21116
- 21116:21116/udp
- 21117:21117
- 21118:21118
- 21119:21119
image: rustdesk/rustdesk-server-s6:latest
environment:
- "RELAY=<PUBLIC_IP>:21117"
- "ENCRYPTED_ONLY=1"
volumes:
- ./data:/data
restart: unless-stoppedwhich is the second one you mentioned (#54 (comment)), without the |
|
I tried both configurations with both Clients at the same network and an external rustdesk server: With Exposing ports and without |
|
Recap:
Am I correct? |
It was internal server, clients on same network But the rest is correct. |
ok, noted I can safely say that:
This is somewhat strange: it's not a slow connection... Maybe some firewall rules? an aggressive IDS/IPS? |
There are no special rules, so I would say that there is no problem.
I did another test with the RustDesk app on my mobile phone over LTE. I tried to connect to the network with the 2 clients and to the network with the remote client. On both it takes 20-30 seconds or it times out.
Aggressive in which way? I tried it with 3 different providers: |
|
@KreativeKrise some one reported old version works well with out big delay. Could you please try out |
|
Of course, with pleasure. But unfortunately, I only get the message |
|
|
Client1 and Client2 on different networks: Client1 and Client2 on same network: |
|
Thanks |
|
@KreativeKrise Could you help test rustdesk/rustdesk-server:1.1.6? |
|
The connection time is identical to version 1.0.0. |
|
Thanks~ |
|
1.1.6 hugely improved my connection times (nearly instant instead of ~30 seconds). I'm connecting to a Ubuntu 18.04 machine from MacOS. Server is running in docker. |
If I want to connect to a remote system it takes 20-30 seconds. But mostly it timeouts after 30 seconds. Both systems are running Windows 10 with RustDesk 1.1.9.
RustDesk Server was installed via docker and docker-compose on a public available server.
I used this docker-compose config:
but I also tried:
Ports 21115-21119/tcp and 21116/udp for incoming connections are open.
Any idea what could cause this long time it needs to connect?
The text was updated successfully, but these errors were encountered: