Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
10.0.13
Description
Just been bitten by this:
[root@myserver ~]# setup_help_tables
|
ERROR 1406 (22001) at line 76: Data too long for column 'url' at row 1
|
[root@myserver ~]# cat /path/to/setup_help_tables
|
#!/bin/sh
|
#
|
# Update the help tables, but ensure we don't write to any binlogs
|
# as this may break downstream slaves.
|
#
|
|
|
(
|
echo "set sql_log_bin = 0;"
|
cat /usr/share/mysql/fill_help_tables.sql
|
) | mysql mysql
|
[root@myserver ~]# rpm -qa | grep MariaDB-server
|
MariaDB-server-10.0.13-1.el6.x86_64
|
[root@myserver ~]#
|
Please look at:
http://bugs.mysql.com/bug.php?id=69564 Please disable writing to binlogs in fill_help_tables.sql
http://bugs.mysql.com/bug.php?id=61520 help-tableformat doesn't match help-files
I reported the first bug to the upstream source over a year ago, and it was fixed some time later. It is frustrating to see the bug reappear here.
Note a big deal but I was looking for online help about multi-source replication commands and this was missing. The patches made by Oracle of course fix this issue so should be easy to apply.
Attachments
Activity
There are 2 issues:
(1) a column size problem
(2) a don't write to binlogs while loading the data problem.
The column size problem's fixed in 5.6.15:
[smudd@myhost ~]$ rpm -qa | grep MySQL-server
|
MySQL-server-5.6.15-1.el6.x86_64
|
[smudd@myhost ~]$ sudo su -
|
[root@myhost ~]# mysql mysql < /usr/share/mysql/fill_help_tables.sql
|
[root@myhost ~]#
|
What's not fixed in 5.6.15 but later is the "don't write to binlogs".
That may have an issue of pushing down help files to downstream slaves which are not at the same
version as the master, and this may break replication but may also leave invalid help information on the downstream server(s).
Don't confuse the two issues which are mentioned here.
However, if you're going to fix one problem fix them both.
already done both.
Strangely enough, the column size problem was said to be fixed in 5.5.36 (see the bug report) and we, of course, have the fix both in 5.5 and in 10.0 ever since. But that fix only fixed tables as they were created, mysql_fix_privilege_tables was not changed, so upgrades were not covered by the fix. There was a second fix, for mysql_fix_privilege_tables, but only in 5.6. I won't even try to understand the reasoning behind that.
It was apparently fixed in 5.7.5 DMR. We don't merge from 5.7 yet. But I can cherry-pick this fix for you.