Root directory filling space quickly

[[email protected]]

Hi Everyone,

I have just downloaded the Zabbix monitoring tool on centOS8(VM host) via Zabbix packages with MySQL and Apache server. Everything is running very smoothly except:
1. The disk space is filling up very quickly. The screenshot below covers a few hours of information, and it has already filled up the root directory from 78% to 91%. At this fillup rate, no amount of disk space is sufficient.
2. Is the MySQL database filling up this space with all the information it's gathering, or is it something else? I allocated 150GB of disk space and thought it would be enough for this single application.
3. How much disk space does it needs, and is there a way to free up space by deleting unnecessary data.

[cyber]

150G divided into 2 parts... / and /home... You have one partition.. 70G mounted to / All your data is under that one mountpoint, not just your DB, but everything else aswell, like logs, your whole "linux", etc.. , It will fill up... Even without the help of zabbix..
But if its just an empty box, not (yet) monitoring many hosts, then your DB should not grow too much (I have one dev box, with quite lot of hosts, but very little data polled, which has 27G DB...) ... You should dig a bit more into that / partition and see, what exactly grows there...

[[email protected]]

The issue in fact is caused by Inodes over-usage. I wonder if removing unnecessary files such as old log files, and extra files can bring this number lower and fix the issue. Is there a permanent solution to this issue?

[ripperSK]

your questions are not very zabbix related - more in the general Linux administration topic.

If you are using all your inodes - this indicates there are many files created on / (rootfs) and deleting the non-wanted ones will decrease the inode usage.

Permanent fix ?

for the system: find out what is generating these files and if possible stop it.. if not possible - consider re-creating rootfs file system with more inodes allocated (this is not trivial - please do not try to re-create it without proper backup and on a running system)
for you: try to learn more about Linux Administration - many handbooks/websites/tutorials/trainigs are available (lot of these are for free).

Asking zabbix community how to fix (most likely) non-zabbix related issues will not help you in the long run.

P.S.: As indicated by cyber the OS partitioning is not done properly and also installation of database on a shared partition is not recommended.

[tim.mooney]

The original poster is using the pre-configured Zabbix VM appliance. They have little to no control over partitioning, inode density, etc.

I'm going to guess that the original poster tried the VM appliance because they wanted something that was mostly pre-configured and would be easy to use, without having to be an expert in Linux.

Unfortunately, the Zabbix appliance seems to be more intended as a "proof of concept" or a learning environment, not a pre-configured environment that's ready for enterprise use. The docker images seem like they're a better choice for a possible enterprise deployment, but they require Docker knowledge, not VM appliance knowledge.

I'm a little puzzled about what would be running that would be creating so many files on the root volume. ripperSK's advice about tracking that down and stopping it is good advice, but I suspect it's going to be a tricky task for anyone that would choose the VM appliance over other installation methods.