"unexpected EOF" at "recursively searching for replication master" phase

On MacOS 10.14.4

```
$ gh-ost \
--cut-over=default \
--exact-rowcount \
--concurrent-rowcount \
--serve-socket-file=./gh-ost.sock \
--postpone-cut-over-flag-file=./gh-ost.cut-over \
--assume-rbr \
--verbose \
--database= \
--table= \
--alter= \
--host=localhost \
--allow-on-master \
--max-load=Threads_running=4 \
--critical-load=Threads_running=10 \
--default-retries=0 \
--chunk-size=10 \
--max-lag-millis=1500 \
--conf=mysql.conf \
--debug
2019-07-01 21:12:33 INFO starting gh-ost 1.0.48
2019-07-01 21:12:33 INFO Migrating ``.``
2019-07-01 21:12:33 INFO connection validated on localhost:3306
2019-07-01 21:12:33 INFO User has REPLICATION CLIENT, REPLICATION SLAVE privileges, and has ALL privileges on `hyperbase_app`.*
2019-07-01 21:12:33 INFO binary logs validated on localhost:3306
2019-07-01 21:12:33 INFO Inspector initiated on kannan:3306, version 5.6.41-log
2019-07-01 21:12:33 INFO Table found. Engine=InnoDB
2019-07-01 21:12:33 DEBUG Estimated number of rows via STATUS: 6
2019-07-01 21:12:33 DEBUG Validated no foreign keys exist on table
2019-07-01 21:12:34 DEBUG Validated no triggers exist on table
2019-07-01 21:12:34 INFO Estimated number of rows via EXPLAIN: 6
2019-07-01 21:12:34 DEBUG Potential unique keys in : []
2019-07-01 21:12:34 INFO Recursively searching for replication master
2019-07-01 21:12:34 DEBUG Looking for master on localhost:3306
[MySQL] 2019/07/01 21:12:34 packets.go:32: unexpected EOF
[MySQL] 2019/07/01 21:12:34 packets.go:118: write tcp [::1]:52797->[::1]:3306: write: broken pipe
[MySQL] 2019/07/01 21:12:34 packets.go:32: unexpected EOF
[MySQL] 2019/07/01 21:12:34 packets.go:118: write tcp [::1]:52798->[::1]:3306: write: broken pipe
[MySQL] 2019/07/01 21:12:34 packets.go:32: unexpected EOF
2019-07-01 21:12:34 ERROR driver: bad connection
2019-07-01 21:12:34 INFO Tearing down inspector
2019-07-01 21:12:34 FATAL driver: bad connection
```

Any idea how to debug this?

Riproduci il comando gh-ost fornito su macOS 10.14.4 ed esamina i log relativi alla fase “Recursively searching for replication master”. Traccia il fallimento della connessione indicato da packets.go e dalla broken pipe, quindi conferma la causa e documenta o convalida una correzione approvata da un maintainer.

Scritto dal modello di indicizzazione a partire dal testo della issue.