Im curious what others have experienced with SD Card life, especially on mast installations. I’ve used Pi’s in the past for various solutions, but found in day to day applications the SD card lasts maybe 4-6 months before you get errors.
I know the MMDVM Pi-Star throws it into read only mode, does the satnogs image do anything special?
What we try is to use memory (RAM) for read/write operations for observations, in the first days we had used sd-cards we had several failures. As far as I can remember after using RAM we haven’t seen a lot sd-cards failures. I would say that there are station that are very active without changing sd-card for at least one year now.
Thats good to know… I had 910 running on a 120Gb SSD but felt its a tiny bit overkill considering everything just gets uploaded to the cloud anyway. Upgrading to the new Buster release currently and using a Sandisk 16Gb Class 10
For our reference stations we have been using the recommended SD cards from this review of SD cards for RPi for 2019. With tens of thousands of observations without failing I can pretty safely say that those are good choices.
Redis was there for facilitating interactions with the web UI (specifically for sockets.io)
Now that the web UI is no longer a part of the satnogs-client (and will be a separate package on its own) there is no need for Redis.
Indeed, satnogs-client uses the WebSocket library gevent-websocket and the use of Redis is appropriate. Therefore, after updating satnogs-client, I still installed the Redis and specified the mandatory use of redis-server.service in the /etc/systemd/system/satnogs-client.service .
It doesn’t use websockets since version 1.0. If you have updated then you don’t need the Redis server. In fact, if you are using satnogs-setup (thus Ansible), it is automatically removed.