dsteinkopf / backup-all-mysql Goto Github PK

on strangle only some of my instances I get this error:

mysqldump: Couldn't execute 'SHOW FIELDS FROM host_summary': View 'sys.host_summary' references invalid table(s) or column(s) or function(s) or definer/invoker of view lack rights to u se them (1356)

the mysql servers that are backed up are mysql 5.7 on ubuntu 18.04.

google says: https://www.osso.nl/blog/mysql-sys-schema-mysqldump-failure/

could you have a look?

Hello,

it possible to modify the script for incremental save ? currently every time is script is running its deleting previous save, can you make a options for saving the file with the current date ? Something like : $DBDUMPSDIR/mysqldump_$db_YYYY_MM_DD_HH_MM.sql.bz2

Thank

In one of my projects, I have very large tables (several GB size). Locking while the table is being dumped would stop the whole system which the dump takes 30 minutes or more. So I decided (as a workaround instead of a "real online backup") to add MYSQLDUMP_ADD_OPTS=--skip-lock-tables. This has been working for months (or years now).

Currently (I assume since PR #8) the backup again locks the project :-(

I was not yet able to dive into that deeply but I assume the following:

• Before, --skip-lock-tables was added after --lock-tables in the command line. So table locking was disabled.
• Now, --lock-tables is repeatedly added only if not backing up "mysql", so it is the last option in the command line. So table locking is active.

As we have disabled tracing (-> removed set -x - for good reasons) I could not see in the log files if the above idea is correct. How about adding some DEBUG option which logs the effective mysql options?