Question / Help Help w/ disconnect after 1-hour of streaming

[Ben80p]

First off, I've been using OBS for a few years now, and I previously haven't had this problem. I've moved around a lot and took a break from streaming, essentially starting over from the ground up.

I'm running into a problem where my stream cuts out after exactly 1-hour of being live. I recently upgraded to a newer router, however this problem existed even before the upgrade and I've been trying to troubleshoot it on my own for roughly 2-weeks. What's worse, when the stream cuts out it also cuts off my network connection and could take a minute or two to reconnect. I've reinstalled OBS and adjusted my bitrate to lower CPU and network usage, but no dice.

Apart from logs, my connection is as followed:
~220Mbps Down, ~35Mbps Up, ~10ms Latency

 

[Narcogen]

Are you running more than one copy of OBS?

16:15:03.068: ================================
16:15:03.068: Warning: OBS is already running!
16:15:03.068: ================================
16:15:03.068: User is now running multiple instances of OBS!

 

[Narcogen]

19:14:39.024: [rtmp stream: 'simple_stream'] Connecting to RTMP URL rtmp://live-jfk.twitch.tv/app...
19:14:39.025: Could not resolve live-jfk.twitch.tv: No such host is known. (11001)
19:14:39.025: [rtmp stream: 'simple_stream'] Connection to rtmp://live-jfk.twitch.tv/app failed: -2

The log indicates that your computer loses the connection to Twitch because it is no longer able to resolve the DNS address of the ingest server.

 

[Ben80p]

Are you running more than one copy of OBS?

16:15:03.068: ================================
16:15:03.068: Warning: OBS is already running!
16:15:03.068: ================================
16:15:03.068: User is now running multiple instances of OBS!

I can easily explain that. When OBS disconnects, it has a tendency to get stuck when trying to start a new stream. I then have to force close OBS, reopen it, and continue streaming.

 

[Ben80p]

19:14:39.024: [rtmp stream: 'simple_stream'] Connecting to RTMP URL rtmp://live-jfk.twitch.tv/app...
19:14:39.025: Could not resolve live-jfk.twitch.tv: No such host is known. (11001)
19:14:39.025: [rtmp stream: 'simple_stream'] Connection to rtmp://live-jfk.twitch.tv/app failed: -2

The log indicates that your computer loses the connection to Twitch because it is no longer able to resolve the DNS address of the ingest server.

But would that explain a consistent disconnect whenever the stream timer rolls over from 00:59:59 to 01:00:00? It stops streaming and my wifi shows a small warning icon at the same time.

Somewhat related:
I contacted my ISP prior to seeing any replies, and they confirmed it's unrelated to them. The person on the phone also happened to be a streamer and had a thought that a cache is filling up somewhere, causing disconnects on exact intervals when full. We went over hardware specifications, internet speeds, possible lost packets, and nothing seemed related, so something's likely amiss in my OBS Studio settings. I plan on trying to stream through OBS classic and seeing if there's any difference, but changing my ingest server sounds like another thing to try.

 

[Narcogen]

No cache failure I can think of would cause DNS resolution to fail.

OBS stopping its own on a timer would not stop DNS resolution to fail.

Stopping streaming from a hotkey would not cause DNS resolution to fail.

An ISP sending a router a reset signal when it sees a connection has remained nailed up for greater than one hour? Now that would cause a DNS resolution to fail.

Try putting the IP address for your intended ingest server (don't use the auto setting, pick the best from the list) and put it in your hosts file. That would prevent DNS resolution from failing. Of course, if your connection is dropping, that won't matter, but at least then you'd see a different error in the log that might be more helpful, because "cannot resolve host name" in this case just means "the internet went away" and there are very few things that can cause that on precise intervals.

 

[R1CH]

Check your firewall / router DHCP settings. One hour is a typical DHCP lease time, perhaps you aren't able to renew your DHCP lease properly which would cause momentary connectivity loss every time it happens. You can confirm with a cmd prompt, run ipconfig /all and look at the Lease Expires time. If you lose internet at this time, DHCP is your issue.

 

[Ben80p]

No cache failure I can think of would cause DNS resolution to fail.

OBS stopping its own on a timer would not stop DNS resolution to fail.

Stopping streaming from a hotkey would not cause DNS resolution to fail.

An ISP sending a router a reset signal when it sees a connection has remained nailed up for greater than one hour? Now that would cause a DNS resolution to fail.

Try putting the IP address for your intended ingest server (don't use the auto setting, pick the best from the list) and put it in your hosts file. That would prevent DNS resolution from failing. Of course, if your connection is dropping, that won't matter, but at least then you'd see a different error in the log that might be more helpful, because "cannot resolve host name" in this case just means "the internet went away" and there are very few things that can cause that on precise intervals.

Actually, I don't use automatic normally, I usually set to "US East: New York, NY"
I think I may have forgotten to change that back after I did a full reinstall of OBS, but I spent today streaming on "NA: Quebec, Canada" with zero change in situation, still a consistent loss of connection every hour, though it auto-reconnects after maybe a minute so long as I don't touch anything.

I should mention that it failed again at something like 28-minutes into today's stream when I tried using "US East: Ashburn: VA," but that looked more like a full on crash than just a disconnect, not sure if that's relevant. After that, I switched to Quebec and continued streaming.

 

[Ben80p]

Check your firewall / router DHCP settings. One hour is a typical DHCP lease time, perhaps you aren't able to renew your DHCP lease properly which would cause momentary connectivity loss every time it happens. You can confirm with a cmd prompt, run ipconfig /all and look at the Lease Expires time. If you lose internet at this time, DHCP is your issue.

Lease Obtained. . . . . . . . . . : Wednesday, June 5, 2019 7:13:39 PM
Lease Expires . . . . . . . . . . : Wednesday, June 5, 2019 8:13:40 PM
You're exactly right! There's no way I would've caught that on my own, or even thought that could be an issue.

Now, if I may ask, how do I go about resolving that? The unfortunate thing is that I'm not incredibly familiar with troubleshooting networking issues, even less so changing things on my firewall.

EDIT1: It's unlikely to be my router, within the time that this has been happening I've since upgraded to a newer router, both are entirely different makes and models. So I think troubleshooting should focus on a change in firewall.

EDIT2: I did some reading into possible solutions to get around dealing with the lease, but my ISP's router is giving me trouble in reserving an IP address (idea #1) and telling my network adapter to use a specific IP address in IPv4 refuses to connect to the internet afterwards (idea #2).

 

[R1CH]

If you're using anything other than the Windows firewall, try removing it and only using the standard Windows firewall for now. Setting a fixed IP address should also work, it just needs to be in the same subnet. Looking at the ipconfig output again, if for example if your router gave you 192.168.1.212 as an address and 192.168.1.1 as a default gateway, you could set a static IP of 192.168.1.2 (or any other unused number), network mask 255.255.255.0 gateway 192.168.1.1, and DNS to 1.1.1.1 / 8.8.8.8.

 

[Melsoniasty]

I have same issue , 10 min and off stream , same RTMP send error
I try this " https://obsproject.com/forum/thread...riten-rtmp-send-error-10054-117-bytes.121894/ " and finally it works

 

[OldVampireLady]

I followed all of the suggestions here to no avail. My twitch stream kept dropping out after 1 hour exactly. Then I discovered that I had somehow set my output timer to 1 hour so that it would cut every stream out at 1 hour. Once I changed that, everything works fine and no dropped streams so far. Just thought I'd mention it -- an obvious fix, but perhaps easily overlooked...