Dockge eating 100% cpu

[officiallymarky]

⚠️ Please verify that this bug has NOT been reported before.

• I checked and didn't find similar issue

🛡️ Security Policy

• I agree to have read this project Security Policy

Description

I noticed my CPU was pegged at 100% with node, so I did a docker stats and I see dockge is eating 100%+ cpu.

👟 Reproduction steps

installed w/ basic docker compose

👀 Expected behavior

minor cpu usage

😓 Actual Behavior

all the cpu usage

Dockge Version

1.1.1

💻 Operating System and Arch

Ubuntu 22.04

🌐 Browser

N/A

🐋 Docker Version

24.0.7

🟩 NodeJS Version

No response

📝 Relevant log output

I checked logs, there really isn't anything happening, yet it is chewing all the cpu.
One thing that maybe the problem, but not sure, is I am running dockge stack inside the same stack folder, I know in the docs it has it outside of it.  Dockge seems to be confused by this and shows it as inactive.  I am not sure if this is related or not. 

I have a lot of containers running, some very complex and even when it isn't at 100% Cpu (when I restart it), it is using way more cpu than any other container by far.

[AlmightyFrog]

Same here. This is the command ps shows doing the load

/usr/local/bin/node --require /pnpm/global/5/.pnpm/[email protected]/node_modules/tsx/dist/preflight.cjs --import file:///pnpm/global/5/.pnpm/[email protected]/node_modules/tsx/dist/loader.mjs ./backend/index.ts

[becks0815]

Can confirm this. "Only" 10% CPU usage here, but this is on the same level like the transmission demaon, which is pushing 10 MB/sec at the moment, while dockge is idle.

Otherwise: great software, it does what portainer doesn't - show the logfile of containers/stacks

[calvinklink]

CPU pegged at 100%. i5-4590 running on SATAIII SSD, Ubuntu Server 22.04.3.
If I restart dockge it's fine. It seems to happen when restart/stop/start actions are performed, mostly at random. Seems to happen more when containers are stopped.

[louislam]

Need reproduce steps and more info, I cannot reproduce even on my Raspberry PI 4.

I suspected it is related to this: #60

But without reproduce steps, it is hard to debug.

[becks0815]

I can only provide some software info: docker on Ubuntu 22.04, Docker version 24.0.7, build afdd53b. I spotted the issue while looking at dozzle, showing me the CPU usage of all containers, but I have no idea if this is caused by some runtime issues and the program starts fine and without eating the CPU, but it is existing.

If I can run any programs/scripts and paste the output here, just mention what I could do to help.

[calvinklink]

Need reproduce steps and more info, I cannot reproduce even on my Raspberry PI 4.

I suspected it is related to this: #60

But without reproduce steps, it is hard to debug.

I don't know if it will get you any closer, but I have noticed it's particularly with Nextcloud's official container, I think. It's the only container that has a persistent volume which also happens to be on a ZFS raid.

[BossMBGaming]

Need reproduce steps and more info, I cannot reproduce even on my Raspberry PI 4.

I suspected it is related to this: #60

But without reproduce steps, it is hard to debug.

I believe that may be the case as I installed Dockge on my Fedora 38 system and I noticed the same issue. I have the same issue of 100% CPU usage, and it tends to happen after using the console for either the system or individual containers.

You can rule out 100% CPU being attributed to an Ubuntu specific issue as the system I have tested Dockge on runs Fedora. But it may be a x64 platform specific issue, which could be preventing you from recreating it on a Raspberry Pi.

My Fedora 38 system is a laptop with an AMD CPU.

My server is running Ubuntu 22.04.3 and uses an Intel CPU. I will test Dockge on my server and see if the 100% CPU usage issue occurs on Intel as well (Will post an update when I have done so). No one mentioned if the processors they use are AMD or Intel, but it is possible this is a factor. I highly doubt it, but I will still test it out when I have some free time.

[louislam]

@BossMBGaming I actually tested on arm64/armv7/amd64, still cannot reproduce. I think it is not related to hardware or os.

I believe it is related to #60, but I don't have evidence, as I cannot reproduce.

[louislam]

I am able to reproduce by deploy/delete/deploy/delete nginx stacks, using the console and clicking around the settings.

I connected the debugger, but the weird things is that it is so normal in the debugger. I suspected it is related to tsx.

[louislam]

Still digging into it. I think it is actually related to node-pty which is used for connecting to "docker compose up/down/restart"'s console. It will sometimes full load the cpu and don't return any data.

I think #143 #169 are caused by this too.

[R0GGER]

@louislam I've send you an email.

[louislam]

@louislam I've send you an email.

Thank you so much. It is able to reproduce now, but I am still struggling on how to identify the issue.

[louislam]

Maybe a related issue: sagemathinc/cocalc#6963