Informational messages are printed twice in 1.0.14

[Roguelazer]

% gh-ost --database easypost --table rates --alter "MODIFY id BIGINT NOT NULL AUTO_INCREMENT" --allow-master-master
# Migrating `easypost`.`rates`; Ghost table is `easypost`.`_rates_gho`
# Migrating `easypost`.`rates`; Ghost table is `easypost`.`_rates_gho`
# Migrating db13sj:3306; inspecting db15sj:3306; executing on db15sj
# Migrating db13sj:3306; inspecting db15sj:3306; executing on db15sj
# Migration started at Mon Aug 29 23:58:19 +0000 2016
# Migration started at Mon Aug 29 23:58:19 +0000 2016

I initially thought that this was caused by --concurrent-rowcount, but it seems to happen regardless of --concurrent-rowcount with 1.0.14, and, in fact, it happens even with no special args, as shown above. Sometimes the messages are out-of-order, so I assume it's a race between two threads. Doesn't seem to affect actual execution.

[shlomi-noach]

@Roguelazer I also recall seeing that before. But I haven't, in a while. Does that happen consistently? Any particular time when it happens?

Also, are you issuing status commands on the unix-socket?

[shlomi-noach]

@Roguelazer friendly ping

[Roguelazer]

I was not issuing status commands on the socket; it double-prints during startup.