|
cyl3x opened 11 months ago
|
||||||
|
Looks like that JVM thinks there are plenty of memory and is lazy on releasing to system. Please reduce memory available to JVM to see if the situation is better. This can be done by changing |
||||||
|
|
Before I set a memory limit of 6gb in Yesterday I limited the container to 4GB with the Currently the server reports a memory usage of ~150mb, while the actual memory usage is 1.1gb. I have seen this difference many times in the past, but why is there a discrepancy between the server's memory usage report and the actual memory usage? Also, the actual memory usage is slowly creeping up while the report stays the same, which seems a bit odd. |
||||||
|
|
Here is a small history of memory usage over the last 70 days (I do not keep older metrics). With a few exceptions, the memory graph only decreases when I restart/update the server or force garbage collection. Therefore, the server does not seem to free any memory at all.
(because of the squeezed timeline, the initial memory usage looks higher than it really is) |
||||||
|
|
JVM can use up to max memory allocated to it (via JVM has some option to release memory to OS and I will experiment with it later. For now, you may further reduce the memory allocated to JVM via |
||||||
|
|
Okay, that is good to know, I will reduce my memory limit further. Thank you for taking the time to investigate :) |
||||||
|
|
cyl3x changed fields 11 months ago
|
||||||
|
|
Robin Shen changed state to 'Closed' 6 months ago
|
||||||
|
|
Memory usage can now be controlled via environment variable |
| Type |
Question
|
| Priority |
Normal
|
| Assignee |
Hi there, for quite some time I experience a suddenly high memory usage.
A month ago I created #1294 and lately my log was spammed with
io.onedev.commons.utils.ExplicitException: Please login to perform this queryerrors. Both issues I discovered because the memory exploded in my grafana dashboard and I wanted to look for the problem in the server logs.Here are three screenshots from different days when memory usage suddenly exploded, the last one is before I limited my server to a maximum of 6Gb:
I think it's caused by the server logger or the database backup at 1am (but the memory usage only lines up sometimes), but to be honest I have no idea.
There is no traffic on the server (as my grafana statistics for in- and outbound traffic is zero), no scheduled build job is running or something like this.
The server itself reports a very low memory usage (real one is 5.12Gb):
My server is mostly idle and therefore I want to eliminate the high memory usage. Thanks for your help :)