Cannot subscribe to podcasts

[Jefronty]

I recently installed version 0.9.0 of pinepods but I am unable to add any podcasts. I've tried uploading an OPML file, manually adding a URL or clicking the (+) in search results. Chrome's developer tools network tab shows the /api/data/add_podcast has a 502 status as does /api/tasks/active?user_id=2 afterwards. Regardless of client UI: mobile, desktop, or web browser, I am unable to add/subscribe to any podcast. I do have debug_mode enabled and the Docker terminal shows the following when I attempt to add a podcast feed:

getentropy failed.                                                                                                  
2026-07-08T12:50:19.781118Z  INFO pinepods_api: Ὠ0 Starting PinePods Rust API...                                   
2026-07-08T12:50:19.782631Z  INFO pinepods_api: Starting PinePods Rust API                                          
2026-07-08T12:50:19.783954Z  INFO pinepods_api: Configuration loaded                                                
2026-07-08T12:50:19.784992Z  INFO pinepods_api: Database config: host=*********, port=3306, user=***********, db=************, type=mariadb                                                                                                 
2026-07-08T12:50:19.790530Z  INFO pinepods_api: Database pool initialized                                           
2026-07-08T12:50:19.794229Z  INFO pinepods_api::redis_client: Successfully connected to Redis/Valkey                
2026-07-08T12:50:19.795594Z  INFO pinepods_api: Redis/Valkey client initialized                                     
2026-07-08T12:50:19.797012Z  INFO pinepods_api: Task management system initialized                                  
2026-07-08T12:50:19.878185Z  INFO pinepods_api: ὕ2 Initializing background task scheduler...                       
2026-07-08T12:50:19.880342Z  INFO pinepods_api::services::scheduler: ὕ2 Starting background task scheduler...      
2026-07-08T12:50:19.881835Z  INFO pinepods_api::services::scheduler: Ὄ5 Podcast refresh schedule: 0 */30 * * * *   

The image I downloaded from hub.docker.com is 0.9.0-amd64
I've verified that the DB credentials used are able to write to the tables, but the same credentials are the ones which created the tables so I didn't really think that was the issue.

[conorbronsdon]

Trying to narrow this down (went to try and solve but couldn't identify).

502 clue: A SQL/insert error in the add path would surface as a 500, not a 502, and the add-podcast INSERT looks dialect-correct for MariaDB (plain parameterized insert + last_insert_id(), no RETURNING/ON CONFLICT).

• A 502 should mean nginx didn't get a valid response from the API: either the API process errored/reset mid-request or it's restarting. The fact that /api/tasks/active also 502s points the same way, since that endpoint has no path that would return a 502 on its own.

Ask: Could you run with DEBUG_MODE=true and grab the container/horust log from the moment the add fails (not just the startup lines)?

Specifically I'm wondering if pinepods-api panics or restarts right then. Claude noticed in your startup output: getentropy failed. The TLS side recovers from that, but the task system calls Uuid::new_v4() when it schedules the episode-fetch after an add, and if getrandom is blocked in your container (seccomp/older kernel) that call panics, which would fit both the add failure and the tasks/active 502, on any database. The add-time log should tell us which it is.

[Jefronty]

@conorbronsdon I do have DEBUG_MODE=true and I included the output from the time it fails. The first line says `getentropy failed.' and then the container seems to reboot. It cannot fetch an episode yet, because no podcasts are successfully added to the Podcasts table in the database. When I tested the DB insert I didn't use a valid podcast URL, I just made sure an insert could be made with the credentials used by pinepods. It was successful, then I cleared it out.

[conorbronsdon]

Here's my take:

• Your manual insert with the pinepods credentials working tells us the database and the add query are seemingly fine. The getentropy failed right before the restart = cause. Something in the Rust stack calls getrandom during the add, and when the container can't get entropy that call panics and takes the API process down, which is also why tasks/active 502s and why the podcast never lands (the insert rolls back). It's showing at the top of your startup logs too, so entropy's blocked from the moment the container starts, not just at add time.
• Likely cause: the container runtime blocking the getrandom syscall, rather than pinepods itself. The official image is Alpine (musl), and Alpine containers (according to Claude) are a known trigger for this getentropy failure when an old host seccomp profile blocks getrandom, so it's probably the host side (its Docker/libseccomp), not the image.

To narrow it: what's the host (a regular distro, or a NAS like Synology/Unraid), and roughly how old is the Docker install? Old Docker or libseccomp is the usual reason getrandom gets blocked. Quickest test: run the container with --security-opt seccomp=unconfined and try adding a podcast. If that works, I think it's the seccomp profile, and updating Docker and libseccomp on the host sorts it.

[Jefronty]

It may be related to the host system, but the previous version of Pinepods (0.8.2) worked on the same system. It is a Synology NAS. I am using Synology's Docker package to edit and manage containers and don't see a way to add additional arguments like --security-opt seccomp=unconfined I'll look into how I can do that. The version of Docker is 20.10.3

[madeofpendletonwool]

Your version of docker is from Feb 2021. I would highly recommend seeing if you can update that. Not only would it likely fix the issue but also fix an absolute entourage of vulnerabilities.