Server Ram utilization

I have been using ERPNext for last 2 years on a VPS with 3GB ram. Whereas initially the requirement was okay but progressively with updates, we are finding that the requirement of RAM has been increasing and with version 8 the system is crashing very often due to mysql crashes. I have collected some RAM utilization data after the reboot. Being a Saturday, the load on the server is low, but the memory utilization is increasing every minute (Memory Leaks?
:thinking: )

I request you to kindly address the issue.



What do you mean by utilization?


Here is what our current active server shows:

[frappe@frappe ~]$ free -m
              total        used        free      shared  buff/cache   available
Mem:          15849        6099         556           8        9193        9438
Swap:          1023         846         177

This is to be expected. A properly configured MariaDB server will behave this way.

Sounds like your use of the system is increasing. As you enter more records, there is more to process - hence larger load.

Add another GB of RAM to your server.

Hi @Amit_Saxena,

seems very similar to an issue we’re currently dealing with. Did you eventually resolve this?

Kind regards,

No, have not been able to figure out a proper solution to the problem. Currently I have a cron job to restart the service every hour. Although not a very neat or a permanent solution but seems to reduce the problem a bit as we still start facing problem after few minutes of restart. Also the problem seem to keep increasing with new releases.


Hi @Amit_Saxena

I eventually decided to jump ship! I setup a fresh server with Ubuntu 16.04 and then migrated my sites using backup/restore

If it’s an option for you, I would totally recommend it unless of course if your usage has exceeded your current capacity in which case you might just need more Ram. Performance is a lot better now


Thanks Wale for your feedback. Is your fresh server a droplet or have you setup a dedicated server?


Hi @Amit_Saxena

It’s a droplet (actually an instance on aws). It’s pretty much identical to what I had originally. I however observed that the issue became much worse after I upgraded the OS on my old server to Ubuntu 16.04 (from 14.04)

That’s why I eventually decided to try a fresh instance and it worked out fine. It’s much better than it was even before I upgraded to 16.04 (on the old server)