Issue with backups (Wordpress & Defender plugin)

namzzz · July 27, 2020, 8:40am

Hello.

I have an issue with backup. Everything was working fine, but today i need to download last backup archive and i’ve noticed, that i have old one from 2020-07-15, but today is 07-27) Nothing on server was changed or updated during that period. I’ve started to dig into logs.

Logs

> /var/log/hestia/backup.log
> 
> ================================
> admin
> --------------------------------
> 
> -- SYSTEM --
> 2020-07-15 05:10:01 admin.conf
> 2020-07-15 05:10:01 stats.log
> 2020-07-15 05:10:01 history.log
> 2020-07-15 05:10:01 pam
> 
> 
> -- WEB --
> 2020-07-15 05:10:01 somedomain.com
> 2020-07-15 05:10:09 *** 1 domain ***
> 
> -- DNS --
> 2020-07-15 05:10:09 *** 0 domains ***
> 
> -- DB --
> 2020-07-15 05:10:09 admin_somedomain (mysql)
> 2020-07-15 05:10:10 *** 1 database ***
> 
> -- CRON --
> 2020-07-15 05:10:10 *** 11 jobs ***
> 
> -- User Dir --
> 2020-07-15 05:10:10 adding .bash_logout
> 2020-07-15 05:10:10 adding .bashrc
> 2020-07-15 05:10:10 adding .cache
> 2020-07-15 05:10:10 adding .composer
> 2020-07-15 05:10:10 adding .config
> 2020-07-15 05:10:10 adding .local
> 2020-07-15 05:10:10 adding .profile
> 2020-07-15 05:10:10 adding .ssh
> 2020-07-15 05:10:10 *** 8 directories ***
> 
> -- SUMMARY --
> 2020-07-15 05:10:10 Rotated: 2020-07-14_05-10-11
> 2020-07-15 05:10:10 Local: /backup/admin.2020-07-15_05-10-10.tar
> 2020-07-15 05:10:10 Size: 55 MB
> 2020-07-15 05:10:10 Runtime: 1 minute
> 
> --------------------------------
> 
> 
> ================================
> admin
> --------------------------------
> 
> -- SYSTEM --
> 2020-07-16 05:10:01 admin.conf
> 2020-07-16 05:10:01 stats.log
> 2020-07-16 05:10:01 history.log
> 2020-07-16 05:10:01 pam
> 
> 
> -- WEB --
> 2020-07-16 05:10:01 somedomain.com
> 2020-07-16 05:10:10 *** 1 domain ***
> 
> -- DNS --
> 2020-07-16 05:10:10 *** 0 domains ***
> 
> -- DB --
> 2020-07-16 05:10:10 admin_somedomain (mysql)
> Error: dump admin_somedomain failed
> 
> --------------------------------

SQL dump problem repeats in log till today. I’ve deleted backup archive from CP today and tried to create a new one. The system still trying to make it with no luck( New stuff in log below:

> ================================
> admin
> --------------------------------
> 
> -- SYSTEM --
> 2020-07-27 05:10:02 admin.conf
> 2020-07-27 05:10:02 stats.log
> 2020-07-27 05:10:02 history.log
> 2020-07-27 05:10:02 pam
> 
> 
> -- WEB --
> 2020-07-27 05:10:02 somedomain.com
> 2020-07-27 05:10:15 *** 1 domain ***
> 
> -- DNS --
> 2020-07-27 05:10:15 *** 0 domains ***
> 
> -- DB --
> 2020-07-27 05:10:15 admin_somedomain (mysql)
> Error: dump admin_somedomain failed
> 
> --------------------------------
> 
> 
> -- SYSTEM --
> 2020-07-27 07:05:01 admin.conf
> 2020-07-27 07:05:01 stats.log
> 2020-07-27 07:05:01 history.log
> 2020-07-27 07:05:01 pam
> 
> 
> -- WEB --
> 2020-07-27 07:05:01 somedomain.com
> 2020-07-27 07:05:09 *** 1 domain ***
> 
> -- DNS --
> 2020-07-27 07:05:09 *** 0 domains ***
> 
> -- DB --
> 2020-07-27 07:05:09 admin_somedomain (mysql)
> sh: 0: getcwd() failed: No such file or directory
> sh: 0: getcwd() failed: No such file or directory
> shell-init: error retrieving current directory: getcwd: cannot access parent directories: No such file or directory
> sh: 0: getcwd() failed: No such file or directory
> exim: can't get the current working directory: No such file or directory
> Error: dump admin_somedomain failed
> -- SYSTEM --
> 2020-07-27 07:10:01 admin.conf
> 2020-07-27 07:10:01 stats.log
> 2020-07-27 07:10:01 history.log
> 2020-07-27 07:10:01 pam
> 
> 
> -- WEB --
> 2020-07-27 07:10:01 somedomain.com
> 2020-07-27 07:10:09 *** 1 domain ***
> 
> -- DNS --
> 2020-07-27 07:10:09 *** 0 domains ***
> 
> -- DB --
> 2020-07-27 07:10:09 admin_somedomain (mysql)
> sh: 0: getcwd() failed: No such file or directory
> sh: 0: getcwd() failed: No such file or directory
> shell-init: error retrieving current directory: getcwd: cannot access parent directories: No such file or directory
> sh: 0: getcwd() failed: No such file or directory
> exim: can't get the current working directory: No such file or directory
> Error: dump admin_somedomain failed
> -- SYSTEM --
> 2020-07-27 07:15:01 admin.conf
> 2020-07-27 07:15:01 stats.log
> 2020-07-27 07:15:01 history.log
> 2020-07-27 07:15:01 pam
> 
> 
> -- WEB --
> 2020-07-27 07:15:01 somedomain.com
> 2020-07-27 07:15:09 *** 1 domain ***
> 
> -- DNS --
> 2020-07-27 07:15:09 *** 0 domains ***
> 
> -- DB --
> 2020-07-27 07:15:09 admin_somedomain (mysql)
> sh: 0: getcwd() failed: No such file or directory
> sh: 0: getcwd() failed: No such file or directory
> shell-init: error retrieving current directory: getcwd: cannot access parent directories: No such file or directory
> sh: 0: getcwd() failed: No such file or directory
> exim: can't get the current working directory: No such file or directory
> Error: dump admin_somedomain failed

Raphael · July 27, 2020, 10:37am

Probaly a wild guess: did you disabled the backup cronjob under admin user?

namzzz · July 27, 2020, 10:41am

Nope, it’s active. If it has been disabled, system should’t write to logs anything about backups, there are everyday 5:10AM records about failed backup.

Raphael · July 27, 2020, 11:26am

Ah, thanks for pointing me to it. First of all: We do not recommend to host anything under the admin user, please create a new user for it.

To your issue, what happens if you try to manualy dump your database using mysqldump? Looks like there is something blocking it.

namzzz · July 27, 2020, 11:58am

Issue solved. After last successful backup i’ve installed Defender plugin for WordPress, which added prefix to existing DB tables. I’ve disabled it a few moments ago, checked v-backup-user via SSH and created backup via CP interface — works perfect.

Thanks for answers and your time!)

falzo · July 27, 2020, 2:43pm

this is interesting. (new) prefixes on the tables or not shouldn’t matter for being able to simply dump the whole db from command-line.
the whole backup process does not care for what’s inside the user dir and database, so normally it cannot depend on the existence or usage of a plugin or not.

maybe however there was something else off/borked with the database and now you somehow fixed/repaired it while fiddling with it.

namzzz · July 27, 2020, 2:50pm

I’ve got error 1356 when i’ve tried dump db manually via SSH. I suppose that plugin Defender does somthing else with user rights and/or smth, not only adds prefixes in the tables’ names. Maybe i’ll dig in it later)

falzo · July 27, 2020, 2:54pm

thanks for adding that detail…

https://boyter.org/2010/10/mysql-error-error1356-view-references-invalid/

maybe that plugin really is using views and not cleaning it up properly or something happened in between which made it stuck.

mehargags · July 28, 2020, 6:22am

I don’t see how a plugin making changes to “inside of a database” affects “MySQLDump” which externally dumps the DB regardless of what tables/names/data is inside of the DB, unless there is a problem in the database structure. So I’d investigate a bit more and keep an eye for failing backups and logs that would point to the real problem