Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Normal memory usage is around 150Mi, but randomly the entire cluster fail and minio uses 2.1Gi. We don't know why this happens, likely a bug in minio. Add a hard limit to get minio killed if it tries to allocate more than 300Mi. Memory usage after starting the env: $ minikube ssh -p dr1 'top -b -o RES -n 1' | head -15 ... PID USER PR NI VIRT RES %CPU %MEM TIME+ S COMMAND 1332 root 20 0 1416.8m 621.1m 6.7 10.5 1:25.59 S kube-apiserver 11906 167 20 0 2235.8m 481.3m 0.0 8.1 0:50.51 S ceph-mgr 7642 root 20 0 1754.0m 116.4m 0.0 2.0 0:00.77 S minio 731 root 20 0 2323.7m 114.7m 0.0 1.9 3:15.37 S /usr/bin/containerd 7288 dbus 20 0 868.2m 107.2m 0.0 1.8 0:05.94 S /bin/package-server Memory usage after running minio self test 100 times: $ for i in $(seq 100); do addons/minio/test dr1; done ... $ minikube ssh -p dr1 'top -b -o RES -n 1' | head -15 ... PID USER PR NI VIRT RES %CPU %MEM TIME+ S COMMAND 1332 root 20 0 1416.8m 651.5m 6.2 11.0 1:36.90 S kube-apiserver 11906 167 20 0 2235.8m 485.5m 0.0 8.2 0:53.57 S ceph-mgr 7642 root 20 0 1821.6m 209.0m 0.0 3.5 0:02.62 S minio 1331 root 20 0 10.8g 118.5m 6.2 2.0 0:27.75 S etcd 731 root 20 0 2323.7m 115.1m 0.0 1.9 3:16.98 S /usr/bin/containerd Signed-off-by: Nir Soffer <[email protected]>
- Loading branch information