MariaDB Restarting with When Disk Quotas Enabled

General Discussion
1

Staff

I have a problem with my server with virtualmin installed.

I configured the disk quotas for the user and when a user’s quota is reaching the limit the mariadb keeps restarting and changing from the active(running) status to the activating(auto-restart) status.

I saw that when running the service mariadb status command I have a preallocating error but I realize that this only occurs when the user of a site is reaching the quota limit and affecting the service.

I tried to enclose a /etc/systemd/system/mariadb.service.d/foo.conf file with the following lines:

[Service]
Restart=on-failure
RestartSec=30s

But without success.

The mariadb log files and screenshots are attached.

Can you help me understand and solve this please

Arquivo de configuração
Arquivo de configuração1366×768 72 KB

MYSQL Ativo
MYSQL Ativo1366×768 142 KB

MySQL Reiniciando
MySQL Reiniciando1366×768 157 KB

2

not to be insulting, but why not just increase the quota?

3

Because each user needs to have their quota and even increasing it I would probably have to free up more and more space and I would always find mariadb stopped for everyone because of 1.

4

I think this is quite an important issue that needs further thought, especially as it looks like the ‘Hard’ quota type is the default option.

MariaDB struggles with this and the result is a database server that constantly restarts.

The work around is to change the quota type to ‘Soft’:

image
image996×604 41.1 KB

System Settings → Server Templates → Default Settings → Administration User

5

Perhaps I’m missing something, but I can’t actually find how one can view what type of quota a virtual server has.

6

There appears to be some sort of loop in this issue that perpetuates due to the disk quota enforcement. The repeating errors are that there are tables that need to be repaired, then something like:

MariaDB restarts

2024-10-06 20:43:05 15844 [ERROR] InnoDB: preallocating 786432 bytes for file ./over_quota_site/wp_wflogins.ibd failed with error 122

241006 20:43:05 [ERROR] mysqld got signal 11 ;

MariaDB restarts

Interesting!

7

maybe I am missing something, why not give that account/virtual server more quota?

8

The issue here isn’t space - the issue is that MariaDB repeatedly restarts, resulting in continuous disconnection / reconnection:

2024-10-08 18:27:11 12 [ERROR] InnoDB: preallocating 786432 bytes for file ./over_quota_site/wp_wflogins.ibd failed with error 122
241008 18:27:11 [ERROR] mysqld got signal 11 ;

2024-10-08 18:27:20 8 [ERROR] InnoDB: preallocating 786432 bytes for file ./over_quota_site/wp_wflogins.ibd failed with error 122
241008 18:27:20 [ERROR] mysqld got signal 11 ;

2024-10-08 18:27:49 19 [ERROR] InnoDB: preallocating 786432 bytes for file ./over_quota_site/wp_wflogins.ibd failed with error 122
241008 18:27:49 [ERROR] mysqld got signal 11 ;

This has a server wide impact.

9

Virtualmin → Logs and Reports → Disk Usage will tell you exactly how much disk quota a virtual server has been alloted and how it is used for web, mail and database.

10

Not come across this problem on any of my VS and some have quite a range of quotas assigned from 1Gb up to 100Gb. (all with default “hard”) I am not aware of anything having changed. Is this particular to some OS or MariaDB version?

11

This doesn’t specify the quota type.

12

It’s not something I have come across before, but I’ve had it reported to me from several clients. This is the first time I have looked into it in detail.

A forum search pulls up several other reported instances of this happening.

This server is pretty much stock:

SYSTEM INFORMATION
OS type and version Redhat Enterprise Linux 8.10
Usermin version 2.102
Virtualmin version 7.20.2
Theme version 21.20.7
Apache version 2.4.37
MariaDB version 10.3.39
Package updates All installed packages are up to date

Looks like OP is running Debian?

1 Like
13

Just run into this issue again.

Does anyone know of an easy way to change ALL disk quota types to ‘Soft’ on a server?

14

and I have just re-examined your previous posts - I should have done that before and have just noticed that you are a WP system (something I will not touch with a barge pole) So my previous comment is inappropriate.

So the error is WP related (quite possibly/probably a plugin/theme) rebuilding/indexing a big table?

Still not sure why the “Hard” quota could be the problem here.
I thought it had been fixed by setting all quota to “Soft”.

15

WordPress has nothing to do with this.

16

The command line and cycle through each site. I am sorry I cannot give you a specific example, just the idea😃.

Can you push changes via the list virtual server page, in pro apparently you can push PHP changes there.

17

A typical Word Press table name/prefix (or why else name it thus?)

Or to be more precise that is a table (one of many) created/managed by the Wordfence plugin. It was listed as producing an error in your post.

(I am just highlighting it as that is what is breaking not the code in Virtualmin.)

and once again the many '000’s of content users are quite happy with the default “Hard” quotas or adjust them as required on a per VS basis.

18

What I mean is that this issue has nothing to do with WordPress.

I’m not saying that this is an issue with Virtualmin. Please read the thread in detail.

19

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.