Applicable to:
- Plesk for Linux
Symptoms
Plesk administrator receives messages about failed logrotate tasks:
/etc/cron.daily/logrotate:
mysqladmin: connect to server at 'localhost' failed
error: 'Access denied for user 'root'@'localhost' (using password: NO)'
error: error running shared postrotate script for '/var/log/mysql/mysql.log /var/log/mysql/mysql-slow.log /var/log/mysql/mariadb-slow.log /var/log/mysql/error.log '
run-parts: /etc/cron.daily/logrotate exited with return code 1
-
Execution of the command below fails:
# mysqladmin --defaults-file=/etc/mysql/debian.cnf --local flush-error-log
mysqladmin: connect to server at 'localhost' failed
error: 'Access denied for user 'root'@'localhost' (using password: NO)' -
User
rootwith no password is stored in the file/etc/mysql/debian.cnf:[client]
host = localhost
user = root
password =
...
Cause
Plesk bug with ID #PPPM-8915 fixed in:
- Plesk Onyx 17.8 25 March 2019
- Plesk Obsidian 28 August 2018 (fixed as PPP-37427, Linux)
Resolution
In case Plesk updates can not be installed, following workaround can be applied in Debian based operating systems:
-
Connect to the server via SSH
-
Create a backup of the initial configuration file:
# cp -a /etc/mysql/debian.cnf /root/debian.cnf
-
Create MySQL database backup:
# MYSQL_PWD=`cat /etc/psa/.psa.shadow` mysqldump -u admin mysql > "mysql`date +%F_%H.%M`.sql"
-
# plesk db
-
Create MySQL user
rootwith password:GRANT RELOAD, SHUTDOWN, PROCESS, SHOW DATABASES, SUPER, LOCK TABLES ON *.* TO 'root'@'localhost' IDENTIFIED BY 'your_password';
-
Exit MySQL:
exit;
-
Modify
/etc/mysql/debian.cnfand specify user with password created above:[client]
host = localhost
user = root
password = "your_password"
socket = /var/run/mysqld/mysqld.sock
[mysql_upgrade]
host = localhost
user = root
password = "your_password"
socket = /var/run/mysqld/mysqld.sock
basedir = /usr -
Make sure that logrotate works. If the command doesn't show the output then everything works properly:
# mysqladmin --defaults-file=/etc/mysql/debian.cnf --local flush-error-log
Comments
30 comments
Step 5 has typo (asterisk symbols were parsed as bold text):
Should be:
GRANT RELOAD, SHUTDOWN, PROCESS, SHOW DATABASES, SUPER, LOCK TABLES ON *.* TO 'debian-sys-maint'@'localhost' IDENTIFIED BY 'PASSWORD';
Instead of:
GRANT RELOAD, SHUTDOWN, PROCESS, SHOW DATABASES, SUPER, LOCK TABLES ON . TO 'debian-sys-maint'@'localhost' IDENTIFIED BY 'PASSWORD';
Good Day
The command on step 5 for MariaDB should be the following:
GRANT RELOAD, SHUTDOWN, PROCESS, SHOW DATABASES, SUPER, LOCK TABLES ON *.* TO 'debian-sys-maint'@'localhost' IDENTIFIED BY 'PASSWORD';
Hello @Sage and @Duncan, thank you for notice.
The article will be reviewed.
Hi,
I also use Ubuntu 18 with Plesk 17.8.11 and MariaDB 10.2.19. I receive the following message every day.
Do you recommend me to do the same workaround? Or is it another problem with a different root cause?
Lots of greets
logrotate_script: line 2: [: /var/run/mysqld/mysqld.pid: binary operator expected
@Markus Wernecke
Hi! Please check the "/etc/logrotate.d/mysql-server" file, if its content differs from the following, please change it accordingly:
/var/log/mysql/mysql.log /var/log/mysql/mysql-slow.log /var/log/mysql/mariadb-slow.log /var/log/mysql/error.log {
daily
rotate 7
missingok
create 640 mysql adm
compress
sharedscripts
postrotate
test -x /usr/bin/mysqladmin || exit 0
if [ -f `my_print_defaults --mysqld | grep -m 1 -oP "pid-file=\K.+$"` ]; then
# If this fails, check debian.conf!
mysqladmin --defaults-file=/etc/mysql/debian.cnf --local flush-error-log \
flush-engine-log flush-general-log flush-slow-log
fi
endscript
}
Thanks!
Hi @Alisa Kasyanova,
thanks a lot for help. Yes there is a difference: ..."pid-file=\K[^$]+"` ];... I changed it to your configuration. Do I have to apply this by a command? So I wait until script will be executed in a few hours. Lots of greets
@Markus Wernecke
You're welcome!
You may test whether it is ok now by running "/etc/cron.daily/logrotate". If you won't receive the same notification, then it is fine.
Thanks a lot for your help. Sorry, I was not able to answer before. This error message doesn't appear anymore. But then, exactly the same day the error message from above in your tutorial appeared. Now I followed your helpful tutorial from above. Hope it works. Lots of greets and thanks for the great support.
@Markus,
Great to hear this helped.
Thank you for letting us know.
Thanks for the workaround!
Seams to work out for me.
Just one note/question:
Will it be fixed with the next update / does the next update reverse this workaround / or do I have to do the next workaround after the next update?
Hello @Wolfgang,
Glad to hear that workaround has helped.
> Will it be fixed with the next update
The bug will be fixed in the new Plesk version right at release but there is no exact date of release.
Currently, there is no ETA for backport to Plesk 17.8
> does the next update reverse this workaround / or do I have to do the next workaround after the next update?
Plesk update/upgrade does not modify the configuration changes of above workaround.
Hi,
I also use Ubuntu 18 with Plesk 17.8.11 and MariaDB 10.2.19. I receive the following message after I follow your instructions above and as last try this:
mysqladmin --defaults-file=/etc/mysql/debian.cnf --local flush-error-log
I get this ERROR message on screen:
mysqladmin: connect to server at 'localhost' failed
error: 'Access denied for user 'debian-sys-maint'@'localhost' (using password: YES)'
I checked the set PASSWORD (replaced with my password) this way:
mysql -u debian-sys-maint -p
Password: PASSWORD (replaced with my password)
and I can login with debian-sys-maint without trouble in mysql - so I think this part will be ok.
So you can give me any hint what I can do?
Best regards
Hi @Joachim Weisse,
It should work in case you can access mysql with this user manually. Please double-check that there are no extra spaces in the password provided in the file after the password or try playing around with quotes.
I have double checked as I wrote above and yes I can login whith the user debian-sys-maint and password I have given. But logrotate does not work. I will try to make you a litle screen recorded video from...
Hello @Joachim,
Thank you for the comment.
As an additional investigation is required, I would suggest submitting a support request. Usually, this allows to find the cause faster.
Hello,
I've the same problem like Joachim after the workaround above:
Password is correct and I can login manually.
@Marcel Klein
Please check the following article as well: https://support.plesk.com/hc/en-us/articles/213361889 Probably, there could be some password misconfiguration between the database and the file. If this won't help, feel free to submit a request as per https://support.plesk.com/hc/en-us/articles/213608509-How-to-submit-a-request-to-Plesk-support-
@Ivan Postnikov
Any update on issue from Marcel and Joachim?
I have the same issue.
Password is correct and I can login manually.
Hello @Yoshi,
After the conversation in this article, we have not heard from them.
The suggestion was to submit a ticket for Plesk support as an additional investigation was required.
The bug described in this article was fixed for Plesk 17.8.11, mu47, which was released on 25th of March.
Hi,
same problem here, installed updates but the problem remain.
plesk version
Product version: Plesk Onyx 17.8.11 Update #49
Update date: 2019/04/09 06:26
Build date: 2019/04/05 05:45
OS version: Ubuntu 18.04
Revision: 2761d524aa7ba76d89920a04e85b43c2b4172878
Architecture: 64-bit
Wrapper version: 1.2
same error even after reboot
the patch guide with update to database is not longer available, so i followed only the main instructions on the post i read yesterday here.
Thank you to drive me on right direction
Hello,
interesting - the error is gone here (Plesk 17.8.11) now.
I haven't enough time last weeks to make a ticket, that is the only reason why you hear nothing from me. :)
I will take a look at the log rotates next days, so I can give feedback.
Hi @Giuseppe D.,
Thank you for letting us known
Possibly, the cause of the issue is related to something else...I recommend creating a request to Plesk Support. We will be glad to assist you, find the root cause of the issue and share our results here in order to make this article useful for other users who met the same issue and were not able to resolve it by resolution in the article
Hi @Joachim Weisse,
Glad to hear that the issue was resolved!
Thank you for the information
@Nikita Nikushkin
il logged in mysql console with admin and pass copied from /etc/psa/.psa.shadow
on the db there are not root or debian.sys-maint users... so i changed the /et/mysql/debian.cnf with admin as user and pass copied from /etc/psa/.psa.shadow
It's a good way to solve the problem?
@Giuseppe D.
It is better to do the following:
1. Create MySQL user root with password Your_password:
MYSQL_LIN: GRANT RELOAD, SHUTDOWN, PROCESS, SHOW DATABASES, SUPER, LOCK TABLES ON *.* TO 'root'@'localhost' IDENTIFIED BY 'your_password';
2. Modify /etc/mysql/debian.cnf and specify user with password used above:
CONFIG_TEXT: [client]
host = localhost
user = root
password = "your_password"
socket = /var/run/mysqld/mysqld.sock
[mysql_upgrade]
host = localhost
user = root
password = "your_password"
socket = /var/run/mysqld/mysqld.sock
basedir = /usr
3. Execute the following command to make sure that logrotate works. If the command doesn't show the output then everything works properly:
# mysqladmin --defaults-file=/etc/mysql/debian.cnf --local flush-error-log
For me this solution solves the logrotate issue but I am then unable to connect via PHPMyAdmin to individual databases:
MySQL said:
Cannot connect: invalid settings.mysqli_real_connect(): (HY000/1045): Access denied for user 'root'@'localhost' (using password: YES)
This seems odd as the grant is on *.*. Feel I must have missed a step somewhere.
Web Host Edition, Plesk Onyx v17.8.11_build1708180613.11 os_Ubuntu 18.04.
Hello @Peter, Please refer to this article to resolve the issue you faced https://support.plesk.com/hc/en-us/articles/213933965 in case the issue will be persists - submit a request to Plesk support https://support.plesk.com/hc/en-us
hello, I am running
I updated Mariadb to 10.3 using these direction for ubuntu 18
here > https://support.plesk.com/hc/en-us/articles/213403429-How-to-upgrade-MySQL-5-5-to-5-6-5-7-or-MariaDB-5-5-to-10-0-10-1-10-2-on-Linux-
everything seemed to go well and no errors.
this was about 3am, then I received an email around 6:30 am about logrotate
here is the message :
# /etc/cron.daily/logrotate:
mysqladmin: connect to server at 'localhost' failed
error: 'Plugin 'unix_socket' is not loaded'
error: error running shared postrotate script for '/var/log/mysql/mysql.log /var/log/mysql/mysql-slow.log /var/log/mysql/mariadb-slow.log /var/log/mysql/error.log '
run-parts: /etc/cron.daily/logrotate exited with return code 1
so i read this report on older version here and wonder why this error occured
server is fresh install about 3 days ago all I did was add 1 domain and 2 subdomains and update Mariadb to 10.3.25 as mention above.
how to check if this error will occur again or is it ok now ?
thanks,
Spiro
# root@server1:~# mysqladmin --defaults-file=/etc/mysql/debian.cnf --local flush-error-log
mysqladmin: connect to server at 'localhost' failed
error: 'Plugin 'unix_socket' is not loaded'
root@server1:~# mysqladmin --defaults-file=/etc/mysql/debian.cnf --local flush-error-log
mysqladmin: connect to server at 'localhost' failed
error: 'Plugin 'unix_socket' is not loaded'
root@server1:~# /etc/cron.daily/logrotate
logrotate_script: 3: [: /var/run/mysqld/mysqld.pid: unexpected operator
How can I fix this ?
regards,
Spiro
Step 3 has several typos:
Should be:
MYSQL_PWD=`cat /etc/psa/.psa.shadow` mysqldump -u admin mysql > "mysql`date +%F_%H.%M`.sql"
Instead of:
MYSQL_PWD=
at /etc/psa/.psa.shadowmysqldump -u admin mysql > mysqldate +%F_%H.%M.sqlIt took a little digging to figure out how to correct the first statement. I hope this helps someone else.
Please sign in to leave a comment.