FailureDetector's logged period might be incorrect.

[snripa]

FailureDetector component, while receiving messages, attampts to log to which perion of pings that message concerns.

It's using shared field
long period
for that purpose.
Due to the fact that it might be updated in another thread (when in the next ping phase), the log message might be wrong and might use updated period value.