Fix PG::UnableToSend using a reconnect #6
Conversation
|
@amicojeko This should fix our issues, please review |
lib/ruote/postgres/storage.rb
Outdated
| yield | ||
| end | ||
| rescue *CONNECTION_ERRORS => e | ||
| if @retry_on_connection_error && retries < 1 |
There was a problem hiding this comment.
So will retry only 1 time? Can be better to set a max_retries param to retry N times?
There was a problem hiding this comment.
Interesting but In my view it's not needed. The kind of error we are facing it's a network issue that closes the PG connection. With 1 retry it reconnects and it works, if that's not the case it may well be another issue and we just raise the error.
|
@MirkoMignini what about now? |
|
@lleirborras I suggest a param delay (ms) between retries. This way we can test N times for a long time so we will see if is stuck forever or is a temporary network issue |
|
@lleirborras This reconnect didn't solve the problem as quasar, that is using this branch, is still affected. Our rails applications aren't affected by PS:UnableToSend problem but only this gem, I see that rails internally retry one time in case of error but using the postgres reset method, I will try the same! |
|
@lleirborras I added a 0.5 sleep between every retry, I will test in scriptoria (and I will increase also the max retries a little bit) |
No description provided.