Easy CPU and Memory Stressing for Docker and SwarmKit
There's lots of images/repos using stress or stress-ng but I needed something
that didn't need a custom ENTRYPOINT/CMD to work so it was easy to demo with
docker service create, so here we go:
Simply running this image with various tags will get you pre-set cpu/memory stressing output.
| Image:Tag | Command |
|---|---|
bretfisher/stress:latest |
stress --verbose --vm 1 --vm-bytes 256M |
bretfisher/stress:256m |
stress --verbose --vm 1 --vm-bytes 256M |
bretfisher/stress:512m |
stress --verbose --vm 1 --vm-bytes 512M |
bretfisher/stress:1024m |
stress --verbose --vm 1 --vm-bytes 1024M |
bretfisher/stress:2cpu256m |
stress --verbose --vm 2 --vm-bytes 256M |
bretfisher/stress:2cpu512m |
stress --verbose --vm 2 --vm-bytes 512M |
bretfisher/stress:2cpu1024m |
stress --verbose --vm 2 --vm-bytes 1024M |
Example
Notice that the task will crash so quickly that it won't "finish" deployment.
docker service create --limit-memory 200M bretfisher/stress:512m
w5656w2kynqt4xip0b321d91y
overall progress: 0 out of 1 tasks
1/1: ready [======================================> ]
verify: Detected task failure
Operation continuing in background.
Use `docker service ps w5656w2kynqt4xip0b321d91y` to check progress.If I looked at the replica list, you'll see Swarm shutting down out-of-memory tasks and starting new ones
docker service ps w5
ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR PORTS
zkvjxdl7o7ej suspicious_lamport.1 bretfisher/stress:512m linuxkit-025000000001 Shutdown Failed about a minute ago "task: non-zero exit (1)"
xtvjqeh190z2 \_ suspicious_lamport.1 bretfisher/stress:512m linuxkit-025000000001 Shutdown Failed 2 minutes ago "task: non-zero exit (1)"
yv7eoc5c7n2z \_ suspicious_lamport.1 bretfisher/stress:512m linuxkit-025000000001 Shutdown Failed 2 minutes ago "task: non-zero exit (1)"
xyutkbaajjc1 \_ suspicious_lamport.1 bretfisher/stress:512m linuxkit-025000000001 Shutdown Failed 4 minutes ago "task: non-zero exit (1)"If I looked at service logs for just one of those tasks, you'd see it start, consume memory, then crash
docker service logs w5 2>&1 | grep yv7eoc5c7n2z
suspicious_lamport.1.yv7eoc5c7n2z@linuxkit-025000000001 | stress: info: [1] dispatching hogs: 0 cpu, 0 io, 1 vm, 0 hdd
suspicious_lamport.1.yv7eoc5c7n2z@linuxkit-025000000001 | stress: dbug: [1] using backoff sleep of 3000us
suspicious_lamport.1.yv7eoc5c7n2z@linuxkit-025000000001 | stress: dbug: [1] --> hogvm worker 1 [7] forked
suspicious_lamport.1.yv7eoc5c7n2z@linuxkit-025000000001 | stress: dbug: [7] allocating 536870912 bytes ...
suspicious_lamport.1.yv7eoc5c7n2z@linuxkit-025000000001 | stress: dbug: [7] touching bytes in strides of 4096 bytes ...
suspicious_lamport.1.yv7eoc5c7n2z@linuxkit-025000000001 | stress: FAIL: [1] (415) <-- worker 7 got signal 9
suspicious_lamport.1.yv7eoc5c7n2z@linuxkit-025000000001 | stress: WARN: [1] (417) now reaping child worker processes
suspicious_lamport.1.yv7eoc5c7n2z@linuxkit-025000000001 | stress: FAIL: [1] (421) kill error: No such process
suspicious_lamport.1.yv7eoc5c7n2z@linuxkit-025000000001 | stress: FAIL: [1] (451) failed run completed in 1sIf I was watching docker events I'd see it create the container, start it, have a oom event, then die
- ~ container create ~
- ~ network create ~
- ~ container start ~
- ~ container oom ~
- ~ container die ~
- ~ network disconnect ~
- ~ container destroy ~