bdx | from a charm's perspective, what is the best way to inquire about the # of units that exist for the type of application that you are? - e.g. how many of myself are there? | 15:31 |
---|---|---|
bdx | pretty sure there is an all_units floating around somewhere | 15:31 |
bdx | thought I would run it by the pros | 15:31 |
Doctor_Nick | what are some good autoscalers to use with juju? | 17:35 |
bdx | Doctor_Nick: depends on the cloud substrate | 19:39 |
bdx | Doctor_Nick: https://jujucharms.com/charmscaler/ | 19:40 |
Doctor_Nick | bdx: aws, gcloud | 21:35 |
Doctor_Nick | I had looked at that, but it's not free software | 21:36 |
babbageclunk | wallyworld: here's that PR I was telling you about: https://github.com/juju/juju/pull/9180 | 22:32 |
wallyworld | babbageclunk: already looking :-) | 22:32 |
babbageclunk | sweet | 22:32 |
wallyworld | babbageclunk: returning an error from the worker will just restart the worker every 3 seconds and spam the logs. maybe the restart delay needs to be longer in this case? | 22:35 |
babbageclunk | wallyworld: yeah, I was wondering about that - how do I do a longer delay, return a specific error type? | 22:37 |
wallyworld | not sure that we can - i know there's a RetryDelay we can set for the worker itself but I'm not sure off hand how to change that for a specific error | 22:38 |
wallyworld | we'll have to look at the worker code | 22:38 |
babbageclunk | ok, I'll take a look | 22:39 |
babbageclunk | wallyworld: I could just change the worker loop so that it doesn't crash with that error, just logs it and sets retry to something bigger than the default 5 seconds - say 1 minute? | 23:06 |
wallyworld | babbageclunk: i think that works - you mean provide an IsFatal() | 23:07 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!