Rocket.Chat icon indicating copy to clipboard operation
Rocket.Chat copied to clipboard

Rapid Increase in Memory Consumption (400MB in one momemnt)

Open dmonn opened this issue 3 years ago • 4 comments

Description:

Might be related to #25199 and other topics.

CleanShot 2022-06-29 at 15 34 04

In a freshly restarted production machine, I see memory hopping up by over 400MB at times (2x - 3x per day), meaning that the server needs to be restarted every other day.

Node seems to consume most of this memory.

CleanShot 2022-06-29 at 15 36 08

Steps to reproduce:

Run a fairly standard RocketChat setup with Docker – MongoDB one the same server. Using GridFS for static files (migrating soon)

Expected behavior:

Memory usage stays consistent or grows with usage.

Actual behavior:

Random jumps in memory usage that sustain and never go down again

Server Setup Information

163554733-cf343739-0cd0-4d60-a0aa-1b701cb788c1

dmonn avatar Jun 29 '22 13:06 dmonn

Having the same problem with Rocket Chat 4.6.2. Mine is running on Docker. I have to restart it overnight becouse without it the VPS is out of memory. ''' Wersja 4.6.2 ID wdrożenia QRdatT6ePDSSKaAWe Wersja silnika aplikacji 1.31.0 Wersja node'a v14.18.3 Migracja bazy danych 258 (30 czerwca 2022 09:47) MongoDB 5.0.7 / wiredTiger (oplog Włączone) Szczegóły zatwierdzenia HEAD: (e8cb66624) Gałąź: HEAD '''

Kris0x0000 avatar Jun 30 '22 08:06 Kris0x0000

Hi, can you check if this will also happen in latest 4.8.1?

Thanks!

dudanogueira avatar Jul 04 '22 13:07 dudanogueira

Any news about that fix ?

colombani avatar Mar 22 '23 18:03 colombani

I'm not sure if this is a bug or a misconfiguration but it persists on the latest 6.9.2 version (using Docker). I found this issue while researching about it. Mine is not an overnight issue but the memory keeps increasing until the machine gets frozen What do you think? Any previous related issues?

ucekmez avatar Jun 27 '24 13:06 ucekmez

Hi, same issue here with latest stable release - running for few weeks until all memory is consumed, then mongo is killed by oom. Workaround i am using for now is restarting every day the container during the night.

Version
6.9.2
Deployment ID
Bb79hnaiR5gtFSTDP
Cloud Workspace ID
6297ba771117930001c4702d
Apps Engine Version
1.42.2
Node Version
v14.21.3
Database Migration
304 (June 25, 2024 5:04 PM)
MongoDB
5.0.20 / wiredTiger (oplog Enabled)
Commit Details
HEAD: (f9a125f75)
Branch: HEAD
Release 6.9.2

JanMikes avatar Jul 13 '24 16:07 JanMikes