mastodon 4.4.1

querkmachine

sean Alright, I've managed to complete all the steps now. A few times it did output the same errors that @mergy was having in this comment, however, and it seems like Redis is also failing to start when running any of the start/restart commands for me as well.

I've spun it up manually by running redis-server /home/user/apps/my_social/mastodon/redis.conf and things seem happy for now.

querkmachine

Something definitely seems unhappy now. The Redis process seems to die every hour or two and requires manually restarting.

querkmachine

querkmachine It indeed seems that something Redis or Mastodon is doing is causing RAM usage to spike, which leads to the server killing the process and supervisord seemingly being unable to restart it.

Is there a way of throttling that, perhaps?

sean

querkmachine Please try adding the following line to supervisor.conf at the beginning of the environment settings (just above RAILS_ENV) for program:puma and program:sidekiq (ie you will add the line in 2 different places):

LD_PRELOAD=/usr/lib64/libjemalloc.so.1,

After you've modified the file, stop the app then kill supervisor like so:

pkill -f supervisor

... then restart the app.

querkmachine

sean Back up and running for the time being, will keep an eye on it to see if it lasts!

Thanks for all the help at weird hours of the day!

querkmachine

querkmachine Well it managed about three days before another RAM spike briefly took it 0.07 MB over the limit and seemingly killed supervisord.

Inconvenient, but at least manageable, and the restart scripts work to get it going again.

« Previous Page