Redis brain hubot script is broken #14

New Issue

2019-02-25T03:39:00Z

raucao commented

2019-02-25 03:39:00 +00:00

Doesn't remember anything anymore. At least data is lost after restarts of the bot.

raucao commented

2019-02-25 04:05:39 +00:00

I looked into it for 10 minutes, but couldn't find anything that looked wrong. Redis is running, and the config is there. This didn't happen from the beginning, so something must have changed at some point.

greg commented

2019-02-25 16:31:16 +00:00

I have also looked at it and can't find anything wrong. I have temporarily enabled debug logs for hal, the brain data is loaded successfully. The last saved word for the plusplus script is pbs, set by bkero on 2019-02-08T18:49:44.895Z

greg commented

2019-02-25 17:07:43 +00:00

Found something, another hubot process appears to be writing to the same Redis key, looking into it!

greg commented

2019-02-25 17:10:27 +00:00

Got it, it's botka. It doesn't have its own REDIS_URL set, unlike the xmpp bots

greg commented

2019-02-25 17:26:27 +00:00

That fixed it! I have set the REDIS_URL for botka so it uses its own database, and removed the --------------- entries from the database (by parsing the JSON with Ruby, editing it and setting the key again).

That fixed it! I have set the REDIS_URL for botka so it uses its own database, and removed the `---------------` entries from the database (by parsing the JSON with Ruby, editing it and setting the key again).

🎉 1

greg referenced this issue from a commit

2019-02-25 17:32:20 +00:00

Set REDIS_URL for botka It was using the same Redis key to write its brain as hal8000 (`hubot:storage`), causing scores to not be persisted to Redis. Right now botka is only saving the online users to the database. It looks like this was only enabled recently, as the last saved score was from Feb 8 Fixes #14

greg closed this issue

2019-02-25 17:32:20 +00:00

raucao commented

2019-02-26 04:03:42 +00:00