electrumx
electrumx copied to clipboard
electrumx crash when bitcoind is restarted + log spamming on INFO and ERROR levels
Today my bitcoind crashed (and restarted automatically, I have script for it), however, electrumx crashed too, shortly after, it didn't handled reconnecting to bitcoind successfully.
log: electrumx.log
Please noticed spam "INFO:Daemon:running normally", "ERROR:Daemon:connection problem - check your daemon is running. Retrying occasionally..." and "ERROR:Daemon:starting up checking blocks. Retrying occasionally..." occuring dozens of times every second. I wouldn't call that "Retrying occasionally". Perhaps we can limit this message to one every 5 seconds?
I am using latest electrumx from git (commit 148df48fca633da0f1d8dcdf8e7f9c1cdcb392fb dated Fri Jul 3 12:17:54 2020 -0500) and bitcoind 0.20.0 on Debian Buster 10 AMD64.
Looks like the pile of retries were from a pile of requests. A single request to the daemon has a backoff routine for retries, but when there are a lot of them, each one of those requests will be retrying. We may be able to fix with a more global throttle.
The crash endured is interesting too. Maybe we need to mark mempool data as dirty prior to resuming mempool sync.