[Bug] Since 1.4.x Multiple timeout issues when connecting / General Instability

[ipat8]

Describe the bug
Since updating to 1.4.0 I've had a host of seemingly random timeouts, crashes, and just instability.

To Reproduce
I don't really have a best way to reproduce, but I found I could most commonly trigger issues by attempting to change any settings like displaying the sidebar, or default visibility, or simply reloading / attempting to load the page while logged in.

The UI loads to the blue background and stalls, eventually the rest of the UI loads in and "Unknown error" pops on the bottom of the page.

I commonly see this in the chrome console:

AuthWrapper.js:15 
i {message: 'timeout of 10000ms exceeded', name: 'AxiosError', code: 'ECONNABORTED', config: {…}, request: XMLHttpRequest}
(anonymous)	@	AuthWrapper.js:15
await in (anonymous)		
(anonymous)	@	AuthWrapper.js:19

GameScanStatus.js:22 [GameScanStatus] shouldPoll: false
GameScanStatus.js:22 [GameScanStatus] shouldPoll: false

Feed.js:85 
i {message: 'timeout of 10000ms exceeded', name: 'AxiosError', code: 'ECONNABORTED', config: {…}, request: XMLHttpRequest}
StatsService.js:9 Failed to fetch folder size: 
i {message: 'timeout of 10000ms exceeded', name: 'AxiosError', code: 'ECONNABORTED', config: {…}, request: XMLHttpRequest}
getFolderSize	@	StatsService.js:9

await in getFolderSize		
(anonymous)	@	Navbar20.js:209
(anonymous)	@	Navbar20.js:216

Navbar20.js:212 Error fetching folder size: 
i {message: 'timeout of 10000ms exceeded', name: 'AxiosError', code: 'ECONNABORTED', config: {…}, request: XMLHttpRequest}
(anonymous)	@	Navbar20.js:212
await in (anonymous)		
(anonymous)	@	Navbar20.js:216

Expected behavior
General stability that was present in prior version 1.3.2

Screenshots

Desktop (please complete the following information):

• All browsers, + mobile

Additional context
For reference, I am hosting via a reverse proxy, but that has never been an issue until now. If you need logs etc, I'm happy to provide.

[ShaneIsrael]

Yeah the connections are timing out for some reason.

Do you have a large library of clips? How large in GB's?

I assume you are running it via Docker?

Do you know what the last version you ran that did not have this issue?

[ShaneIsrael]

Also do you have transcoding enabled and if so, are you using cpu transcoding or gpu? And also if so, what sort of hardware do you have hosting Fireshare?

Timeouts like this could be caused due to the system not having any processing power available to process the requests, such as a transcode using all of the allocated cpu power that the container is given.

If you can provide any logs that appear while this happens that would be great as well.

[ipat8]

If you have specific logs in the container I can grab them, or just the output from TrueNAS. I believe I started to see this issue right after the update that added the option to hide the game section, but I remember the game section causing problems as well.

I didn't explicitly enable transcoding, and all of my files are .mp4, h264, and my docker host isn't reporting CPU consumption by the container. I tested this and it does seem that adding media triggers it consistently, and it will then stabilize-ish over time. I don't see any CPU consumption jump on the host.

Looking at the dockerhub listing I would suspect 1.3.2 was the last stable version, being that I go a while without updating my containers.

[ipat8]

Should also add that the library size is 206 videos, but they're all multiple GBs. Library size is probably 1.2T

[glenjoy99]

Having the same issues here as @ipat8 mentioned, however, I have a significantly smaller library. About 18 videos totaling only 13.5GB.

[ShaneIsrael]

@ipat8 Okay, I have an idea of what might be causing it, and likely due to your large library size.

@glenjoy99 In your case though I'm not entirely sure if its related since your library is much smaller and even my own library of ~200+ GB doesn't have this problem. Can you share anything about your hardware specs and whether or not you have transcoding enabled?

I'll try to get a temporary fix out later today, no promises on whether it fixes it but it will be a good test to help me figure out what the problem actually is. Will respond here when I have a build ready.

[ShaneIsrael]

I know I said later today, but actually an update is ready.

please try the shaneisrael/fireshare:develop docker build. Make sure you force pull it if you already are using that build so that it gets the updates. This dev build has a lot of changes not yet in the production build as well but mainly I want to see if this update I applied resolves the timeout issues you were both seeing.

[glenjoy99]

I am running it on a Raspberry Pi 4B (8GB RAM) with transcoding turned off. Up until the latest release I have had no issues. I unfortunately cannot pull the shaneisrael/fireshare:develop build as I'm getting this error

Regarding the errors, here is the console output that's producing the same blue-screen and "Unknown error" notification at the bottom.

EDIT: Something I have also noticed which may support debugging: When I encounter the "Unknown error" with blue screen, its usually when I first navigate to the main page (i.e. fireshare.xxx.com). Going to the root URL typically redirects to /#/feed which is where the public videos are and thats where the Unknown error pops up. When I navigate elsewhere on the sidebar, such as 'Games', then navigate back to Public Videos the error goes away and my public videos are finally loaded.

[ShaneIsrael]

@glenjoy99 thanks for the info and yeah unfortunately the dev build isn't built for arm :/. Can you try to run version v1.3.2 and does that version have the same issues?

[glenjoy99]

Reverted to 1.3.2. Have not encountered any of those issues. Seems to be stable.

[ShaneIsrael]

@glenjoy99 Okay, if you were using the Steam Grid DB integration could you try the latest version v1.4.0 but disable the Steam Grid DB integration by removing your API key, saving, and restarting fireshare and let me know if that improved anything or not at all.

Just trying to narrow down as best as possible what is causing the issue since I'm not able to recreate the problem on my end. I'm thinking one of the processes is fairly heavy on the cpu and since your on an raspberry pi, you don't have much for cpu power.

So if I can figure out what the likely culprit is, I'll be able to work on fixing it. Unfortunately there's been a lot of updates in the last 2 releases.

[glenjoy99]

Commented out the API key for v1.4 and redeployed. Same issue persists with 'Unknown error' and being met with a blue-screen and no videos. Console output is also the same as what I pasted earlier.