Start9 Server Pure having stability issues

I had good success with everything running for the first few months. I’m currently having trouble where my Start9 Server Pure does not seem to stay up and accessible for long periods. It seems like the first things to go are the web services pages, followed by the base StartOS pages. A reboot usually fixes this. I initially thought it was this issue

but I followed the instructions and reinstalled Bonjour many times and the only thing that seems to help is the reboot.

I do see many successful health checks of all my apps, plus these occasional errors in the OS log:

2023-09-25T14:12:48-05:00 2023-09-25T19:12:48.382884Z ERROR startos::net::vhost: Error in VHostController on port 443: Filesystem I/O Error: timed out
2023-09-25T14:12:48-05:00 2023-09-25T19:12:48.382937Z DEBUG startos::net::vhost: Error { source:
2023-09-25T14:12:48-05:00 0: timed out
2023-09-25T14:12:48-05:00 Location:
2023-09-25T14:12:48-05:00 /rustc/90c541806f23a127002de5b4038be731ba1458ca/library/core/src/convert/mod.rs:727
2023-09-25T14:12:48-05:00 Backtrace omitted. Run with RUST_BACKTRACE=1 environment variable to display it.
2023-09-25T14:12:48-05:00 Run with RUST_BACKTRACE=full to include source snippets., kind: Filesystem, revision: None }
2023-09-25T14:12:48-05:00 2023-09-25T19:12:48.383160Z ERROR startos::db: WebSocket Closed: Network Error: WebSocket protocol error: Connection reset without closing handshake
2023-09-25T14:12:48-05:00 2023-09-25T19:12:48.383183Z DEBUG startos::db: Error { source:
2023-09-25T14:12:48-05:00 0: WebSocket protocol error: Connection reset without closing handshake
2023-09-25T14:12:48-05:00 1: Connection reset without closing handshake
2023-09-25T14:12:48-05:00 Location:
2023-09-25T14:12:48-05:00 /rustc/90c541806f23a127002de5b4038be731ba1458ca/library/core/src/convert/mod.rs:727
2023-09-25T14:12:48-05:00 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ SPANTRACE ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
2023-09-25T14:12:48-05:00 0: startos::db::deal_with_messages
2023-09-25T14:12:48-05:00 at src/db/mod.rs:78
2023-09-25T14:12:48-05:00 1: startos::db::ws_handler
2023-09-25T14:12:48-05:00 at src/db/mod.rs:32
2023-09-25T14:12:48-05:00 Backtrace omitted. Run with RUST_BACKTRACE=1 environment variable to display it.
2023-09-25T14:12:48-05:00 Run with RUST_BACKTRACE=full to include source snippets., kind: Network, revision: None }

How does one set the environment variables like RUST_BACKTRACE? I can SSH into my node, but I don’t see a file with env variables that I should be editing.

Are the errors I pasted above indicative of what the problem might be or should I not be concerned about them? If not those errors, what else should I be looking at?

This looks like a known bug (already fixed for v0345), where the websocket connection gets lost for a few seconds every 5-10min. I might need more details to understand your problem. If the service pages are becoming unavailable, but your main UI is up, then this is indicative of Bonjour issues. With what errors does the main UI become unavailable? Is the server accessible on another device when it becomes unavailable on your Windows machine?

It seems like the services go down first and permanently until I reboot. The main UI eventually stops working as well until a reboot. I have noticed it happen twice in the last month. At the moment, everything seems to be running fine. It seems to take days for this to occur, so I may not have better answer for a few days. I also haven’t confirmed that I lose access from another device, so I will look out for that. If I do still have access via another non-Windows device, then I’ll assume it is Bonjour somehow and post so here.

I haven’t been able to narrow down which errors are occurring around the time when it stops working since I haven’t been watching it closely enough to realize what day it has stopped working. I will have to pay closer attention and get back to you.

I’ll follow up regardless of what happens. I appreciate you taking the time to help me.

Still no problems yet. I’ll update here if I have any more.

Thanks for the updates. I’m going to marked this as resolved for the time being. The symptoms sound very strange and we would need a lot of detail to understand if/what problems are under the hood.

Still no issues these several months later. If there ever was an issue, I’m sure it is fixed with the current version. Thanks for a great product!

4 Likes