Closed Julianoe closed 4 months ago
I'm also seeing large RAM usage. 5GB for Node, with not a lot of activity on the server.
Does this still occur on latest 6.8 or 6.9?
Some indexes for collection 'rocketchat_uploads' could not be created: Index with name: rid_1 already exists with different options
Yes you can delete the index and let Rocket re-create it - did it yesterday:
https://github.com/RocketChat/Rocket.Chat/issues/31876
You may need several restarts to clear them all as new ones may re-appear on restart.
For comparison these are from two similar servers with a dockerised Rocket and native Mongo
720.180 MB /usr/bin/mongod
784.887 MB node
696.180 MB node
768.648 MB /usr/bin/mongod
One of them started life at about 0.32 and been upgraded ever since.
I'd get upgraded to Mongo 5.x but you need to get this fixed first:
WARNING: MongoDB 5.0+ requires a CPU with AVX support, and your current system does not appear to have that! see https://jira.mongodb.org/browse/SERVER-54407
That'll probably mean replacing your server.
Note you can set memory limits in docker if required.
You clearly have a pile of other stuff going on there looking at your logs. Do you have a lot of integragtions?
This issue has been marked as stale because there has been no further activity in the last 10 days. If the issue remains stale for the next 4 days (a total of 14 days with no activity), then it will be assumed that the question has been resolved and the issue will be automatically closed.
This issue was closed because it has been inactive for 14 days since being marked as stale.
Description:
Might be related to several other topics. I feel that my RocketChat instance is using way too much memory for the size and activity there is on it.
Steps to reproduce:
Open several rooms among which room with old history and thousands of messages (11000) from 2017 to 2024.
Expected behavior:
Memory usage stays consistent or grows with usage. When the active user closes a tab, the memory is freed in some way.
Actual behavior:
Simply loading the room, with 1 user only on the server, without sending any messages, makes the memory usage go from 140MB after restarting fresh the database container to skyrocketing and capping at 860MB (22% memory used). Closing the tab (0 active user on the server) makes no difference, the memory usage stays the same until container or server is restarted.
Server Setup Information:
Client Setup Information
Additional context
Activity: 2-6 users not hugely active on a normal day, 50 channels/rooms. Rarely more than 3 users connected at the same time.
Setup I've installed it on a 4Gb RAM server 2 Core Debian 12 (Intel Atom C2338 @ 2x 2.1GHz). I use Docker.
I'm not sure exactly why that could be but it seems to be related to the size of the room. When I open several rooms it does the same too though. 850MB for 3 actives users and a few dozens messages per day seems really excessive and I don't remember it being like that in past years.
For context I host a small instance since 2017.
Using the command
I can't see the mongodb process going from 140MB to 850MB in a few seconds
Relevant logs:
Database logs after restarting:
Channel logs when loading the tab
Docker app container returns this error on start-up, I don't know if it's related: