Conversation

Notices

Embed this notice
silverpill (silverpill@mitra.social)'s status on Friday, 06-Jan-2023 22:32:50 JST silverpill

@japananon Mitra just writes messages to stdout. If you're running it as a systemd service, you can read the log with journalctl -u mitra

In conversation Friday, 06-Jan-2023 22:32:50 JST from mitra.social permalink
- Embed this notice
  silverpill (silverpill@mitra.social)'s status on Wednesday, 01-Feb-2023 11:12:54 JST silverpill
  in reply to
  - チャノさん
  @japananon
  >Federation getting "stuck"
  This just has happened to me. Background tasks stopped working but otherwise the instance was functioning normally. After restart it continued to process all queued messages.
  Could be related to this bug https://codeberg.org/silverpill/mitra/issues/24 (previously, it was fatal)
  In conversation Wednesday, 01-Feb-2023 11:12:54 JST permalink
  Attachments
  1. Domain not in remote thumbnail source whitelist: codeberg.org
    
    Error accepting connection: Too many open files (os error 24)
    
    from silverpill
    
    On rare occasions Mitra stops handling incoming HTTP connections and freezes. The log contains many lines like this: ``` actix_server::accept [ERROR] Error accepting connection: Too many open files (os error 24) ``` Can happen with both Nginx and Caddy. Possibly related: https://github.com/actix/actix-net/issues/351
- Embed this notice
  チャノさん (japananon@mitra.anon-kenkai.com)'s status on Wednesday, 01-Feb-2023 11:12:55 JST チャノさん
  in reply to
  
  @silverpill Ah right.. in that case, nothing unusual in the logs that I've seen yet. My experience of Federation getting "stuck" is only anecdotal, maybe I'm just perceiving things that way. If I happens again I'll make sure to check stdout before restarting anything.
  In terms of the media folder, it's gotten up to about 8.9Gb now. I feel like the growth is slowing down, but if it gets too close to 3/4 full I might change the pruning job to 15 days old instead of 30, see if that improves matters. I figure two weeks of post history is more than enough.
  In conversation Wednesday, 01-Feb-2023 11:12:55 JST permalink
  Attachments
  1. Untitled attachment
- Embed this notice
  silverpill (silverpill@mitra.social)'s status on Monday, 06-Feb-2023 10:44:45 JST silverpill
  in reply to
  - チャノさん
  @japananon In v1.13.0 I added something that could help us find the root cause. Mitra should now throw an error if background task gets stuck (assuming I correctly identified the place where this problem occurs).
  I you don't mind, please disable the cronjob. Let's see what happens
  
  In conversation Monday, 06-Feb-2023 10:44:45 JST permalink
- Embed this notice
  チャノさん (japananon@mitra.anon-kenkai.com)'s status on Monday, 06-Feb-2023 10:44:47 JST チャノさん
  in reply to
  
  @silverpill I'm glad it wasn't just my imagination. My cludge so far has been a cronjob that restarts the mitra service every 12 hours. Since I set that up, the timeline has (so far) remained up to date, with no 'sticking"
  
  In conversation Monday, 06-Feb-2023 10:44:47 JST permalink
- Embed this notice
  silverpill (silverpill@mitra.social)'s status on Thursday, 23-Feb-2023 05:54:55 JST silverpill
  in reply to
  - チャノさん
  @japananon If the error occurs you wouldn't notice it unless you monitor your logs.
  Does journalctl -u mitra -o cat -b | grep ERROR return any results?
  
  In conversation Thursday, 23-Feb-2023 05:54:55 JST permalink
- Embed this notice
  チャノさん (japananon@mitra.anon-kenkai.com)'s status on Thursday, 23-Feb-2023 05:54:56 JST チャノさん
  in reply to
  
  @silverpill I'll run this tonight & keep you appraised of any changes
  
  In conversation Thursday, 23-Feb-2023 05:54:56 JST permalink

Public

Conversation

Notices

Feeds