opencloud-eu
Random very high CPU load while unused
Describe the bug
A few months ago I have setup Opencloud, but haven't used it. It was just running in the background. There is currently only 1 user / 1 file stored. Every few days my serverload would go to 200+ (while it is normally < 2). Running top on my Proxmox node shows that opencloud is hogging the CPU.
If I restart the opencloud container, the load (and resulting temperature) goes back to normal. This is fine for a while and then happens again.
Steps to reproduce
It "just happens" so there is no clear cause but essentially:
- Install Opencloud on LCX => Docker
- Upload 1 file
- Wait a few days
- See the CPU load increase significantly
- Stop the container, see that the load returns to normal
Expected behavior
No CPU overload
Actual behavior
Very high CPU load
Setup
Using the Docker-compose config
Additional context
- My setup is Proxmox => LXC with Alpine => Docker => OpenCloud.
- Seems to be caused by something in 3.5.0, before it was not a problem.
Logs
Some log excerpts:
opencloud-1 | 2025-10-19T13:27:40.830753439Z 2025-10-19T13:27:20Z ERR failed to get consumer error="context deadline exceeded" line=github.com/opencloud-eu/opencloud/services/search/pkg/service/event/service.go:216 service=search
opencloud-1 | 2025-10-19T13:27:41.004398061Z 2025-10-19T13:27:19Z ERR failed to get consumer error="context deadline exceeded" line=github.com/opencloud-eu/opencloud/services/postprocessing/pkg/service/service.go:439 service=postprocessing
opencloud-1 | 2025-10-19T13:27:41.982195238Z 2025-10-19T13:27:37Z ERR registration error for external service eu.opencloud.api.auth-basic error="Failed to store data in bucket '<redacted>': nats: timeout" line=github.com/opencloud-eu/opencloud/pkg/registry/register.go:46 service=auth-basic
opencloud-1 | 2025-10-19T13:56:54.600976977Z 2025-10-19T13:56:48Z WRN Internal subscription on "$JS.API.CONSUMER.INFO.main-queue.search-pull" took too long: 2.630413331s line=github.com/opencloud-eu/opencloud/services/nats/pkg/logging/nats.go:27 service=nats
opencloud-1 | 2025-10-19T13:56:54.600982826Z 2025-10-19T13:56:49Z ERR failed to get consumer error="context deadline exceeded" line=github.com/opencloud-eu/opencloud/services/search/pkg/service/event/service.go:216 service=search
opencloud-1 | 2025-10-19T13:56:54.600985792Z 2025-10-19T13:56:49Z WRN Internal subscription on "$JS.API.CONSUMER.INFO.main-queue.postprocessing-pull" took too long: 2.031193559s line=github.com/opencloud-eu/opencloud/services/nats/pkg/logging/nats.go:27 service=nats
opencloud-1 | 2025-10-19T13:56:54.600988695Z 2025-10-19T13:56:49Z WRN 127.0.0.1:36260 - cid:65 - Readloop processing time: 3.962573872s line=github.com/opencloud-eu/opencloud/services/nats/pkg/logging/nats.go:27 service=nats
opencloud-1 | 2025-10-19T13:56:54.600991571Z 2025-10-19T13:56:49Z ERR registration error for external service eu.opencloud.web.ocm error="Failed to store data in bucket
@syphernl Can you share some more tails about your setup?
Which features do you have deployed? Collabora? Tika? Or is it really just the opencloud container?
Do you have and CPU or memory limits configured in docker or lxc or otherwise?
What kind of system is your setup running on (CPU arch, RAM and so on)
Also, do you see memory usage going up for the opencloud process before, or while the CPU usage is high?
@syphernl Can you share some more tails about your setup?
Which features do you have deployed? Collabora? Tika? Or is it really just the opencloud container?
It is just the opencloud container.
Do you have and CPU or memory limits configured in docker or lxc or otherwise?
Docker has no limits, the LXC has been given 8 cores + 8GB of RAM. No swap. Storage on NVME.
What kind of system is your setup running on (CPU arch, RAM and so on) Latest Proxmox on an Intel i5-13500 with 128GB memory.
Also, do you see memory usage going up for the opencloud process before, or while the CPU usage is high?
I haven't checked the memory usage, just noticed that the load of the entire machine went to 200+ and causes other apps to become unresponsive.
For reproducing we need more info. maybe a CPU trace could help. are you familiar with?
