Slave UUID in replica for gh-ost shows java.lang.ArrayIndexOutOfBoundsException
- Dominant language
- Go
- Stars
- 13.6k
- Forks
- 1.4k
- Avg merge
- 2h 31m
- Merged PRs (30d)
- 4
Description
While using v.1.1.5 (due to compatibility issues of gh-ost's latest version with MySQL 5.7) everything runs fine, but we noticed, after issuing a
```show slave hosts ;```
command in the replica gh-ost uses to read the binlog, that the "Slave_UUID" column shows an error message rather than a proper value
```
Server_id 99999
Host ip-10-0-0-113.ec2.internal
Port 3306
Master_id 1778116519
Slave_UUID
java.lang.ArrayIndexOutOfBoundsException: Index 49 out of bounds for length 49
at org.mariadb.jdbc.internal.com.read.resultset.rowprotocol.TextRowProtocol.setPosition(TextRowProtocol.java:135)
at org.mariadb.jdbc.internal.com.read.resultset.SelectResultSet.checkObjectRange(SelectResultSet.java:635)
at org.mariadb.jdbc.internal.com.read.resultset.SelectResultSet.getString(SelectResultSet.java:943)
in JdbcHelperImpl.getObject(JdbcHelperImpl.java:337)
```
Is this a known issue that is fixed in newer versions? Or maybe a DBMS issue rather than a gh-ost issue?
Again, migrations run fine, we've moved several tables and around 700M rows with no issues so far.
We're using AWS RDS, our setup is as follows:
Master cluster, hosting the master writer and 3 Aurora read replicas
binglog format is MIXED
Engine is 5.7.mysql_aurora.2.11.3
5.7.12-log
Replica cluster, set up manually to replicate via binlog from the master writer, only 1 writer
binlog format is ROW
filtering is OFF (0)
Engine is 5.7.mysql_aurora.2.11.3
5.7.12-log
I'll gladly give more info on the issue if needed.
Contributor guide
Assessment
This issue has not been assessed yet.