Stress-ng: Using pubma sys_admin stress-ng command fails on EKS cluster

[uditgaurav]

trafficstars

Details

• I'm running stress-ng using Pumba on the EKS cluster. After the chaos completion, the Pumba pod gets into the Failed state, and the Pumba pod log gives the following error:

time="2020-10-20T11:43:21Z" level=debug msg="stop stress containers after duration" duration=2m0s
time="2020-10-20T11:43:21Z" level=info msg="stopping container" dryrun=false id=ac9619415b5903f1f0be08293ef22e28d2d647dd476b12a8686390f6024d9501 timeout=5s
time="2020-10-20T11:43:24Z" level=fatal msg="one or more stress test failed: stress-ng failed with error: stress-ng exited with error: \x02\x00\x00\x00\x00\x00\x00\x1ecgroup change of group failed\n"

For complete logs ref: complete_log.txt

[alexei-led]

@uditgaurav I can reproduce this error. It seems you cannot run stress on K8s now. It will take me some time to inspect the failure and fix this issue. Thank you for reporting it!

[PopaRadu]

Hello! Do we have any news about this?

[VamKon]

I have the same issue. Can you please help us out?

[alexei-led]

Still open - need to find time to fix it. Any PR is welcome.

[zhangchl007]

@alexei-led the issue occurred because of cgroup, the dockhack is too old , you must change it , I had already fixed the issue

Please see the commit: https://github.com/alexei-led/stress-ng/pull/3/commits/c7a981db1a1ce91654fcec985ac542d83370578c