I have three master nodes with each 80 GB disk size. Recently I ran into this problem:
Normal Pulling 52s (x2 over 6m17s) kubelet, 192.168.10.37 pulling image "gcr.io/kubeflow-images-public/tensorflow-serving-1.8gpu:latest"
Warning Evicted 8s (x5 over 4m19s) kubelet, 192.168.10.37 The node was low on resource: ephemeral-storage.
–> "The node was low on resource: ephemeral-storage."
The storage on the execution node looks like this:
Filesystem Size Used Available Use% Mounted on
overlay 7.4G 5.2G 1.8G 74% /
tmpfs 3.9G 0 3.9G 0% /dev
tmpfs 3.9G 0 3.9G 0% /sys/fs/cgroup
/dev/vda1 7.4G 5.2G 1.8G 74% /opt
/dev/vda1 7.4G 5.2G 1.8G 74% /mnt
/dev/vda1 7.4G 5.2G 1.8G 74% /media
/dev/vda1 7.4G 5.2G 1.8G 74% /home
none 3.9G 1.5M 3.9G 0% /run
/dev/vda1 7.4G 5.2G 1.8G 74% /etc/resolv.conf
/dev/vda1 7.4G 5.2G 1.8G 74% /etc/selinux
/dev/vda1 7.4G 5.2G 1.8G 74% /etc/logrotate.d
/dev/vda1 7.4G 5.2G 1.8G 74% /usr/lib/modules
devtmpfs 3.9G 0 3.9G 0% /host/dev
shm 64.0M 0 64.0M 0% /host/dev/shm
/dev/vda1 7.4G 5.2G 1.8G 74% /usr/lib/firmware
none 3.9G 1.5M 3.9G 0% /var/run
/dev/vda1 7.4G 5.2G 1.8G 74% /etc/docker
/dev/vda1 7.4G 5.2G 1.8G 74% /usr/sbin/xtables-multi
/dev/vda1 7.4G 5.2G 1.8G 74% /var/log
/dev/vda1 7.4G 5.2G 1.8G 74% /etc/hosts
/dev/vda1 7.4G 5.2G 1.8G 74% /etc/hostname
shm 64.0M 0 64.0M 0% /dev/shm
/dev/vda1 7.4G 5.2G 1.8G 74% /usr/bin/system-docker-runc
/dev/vda1 7.4G 5.2G 1.8G 74% /var/lib/boot2docker
/dev/vda1 7.4G 5.2G 1.8G 74% /var/lib/docker
/dev/vda1 7.4G 5.2G 1.8G 74% /var/lib/kubelet
/dev/vda1 7.4G 5.2G 1.8G 74% /usr/bin/ros
/dev/vda1 7.4G 5.2G 1.8G 74% /var/lib/rancher
/dev/vda1 7.4G 5.2G 1.8G 74% /usr/bin/system-docker
/dev/vda1 7.4G 5.2G 1.8G 74% /usr/share/ros
/dev/vda1 7.4G 5.2G 1.8G 74% /etc/ssl/certs/ca-certificates.crt.rancher
/dev/vda1 7.4G 5.2G 1.8G 74% /var/lib/rancher/conf
/dev/vda1 7.4G 5.2G 1.8G 74% /var/lib/rancher/cache
devtmpfs 3.9G 0 3.9G 0% /dev
shm 64.0M 0 64.0M 0% /dev/shm
/dev/vda1 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/0181228584d6531d794879db05bf1b0c0184ed7a4818cf6403084c19d77ea7a0/merged
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/655a92612d5b43207cb50607577a808065818aa4d6442441d05b6dd55cab3229/merged
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/b0d8200c48b07df410d9f476dc60571ab855e90f4ab1eb7de1082115781b48bb/merged
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/f36e7d814dcb59c5a9a5d15179543f1a370f196dc88269d21a68fb56555a86e4/merged
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/842157b72f9155a86d2e4ee2547807c4a70c06320f5eb6b2ffdb00d2756a2662/merged
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/cee5e99308a13a32ce64fdb853d2853c5805ce1eb71d0c050793ffaf8a000db9/merged
shm 64.0M 0 64.0M 0% /var/lib/docker/containers/6ee5a7ad205bf24f1795fd9374cd4a707887ca2edd6f7e1b4a7698f51361966c/shm
shm 64.0M 0 64.0M 0% /var/lib/docker/containers/79decf02c3a0eb6dd681c8f072f9717c15ba17fcb47d693fcfa1c392b3aef002/shm
shm 64.0M 0 64.0M 0% /var/lib/docker/containers/acc7d374f838256762e03aea4378b73de7a38c97b07af77d62ee01135cc1377b/shm
shm 64.0M 0 64.0M 0% /var/lib/docker/containers/46cb89b550bb1d5394fcbd66d2746f34064fb792a4a6b14d524d4f76a1710f7e/shm
shm 64.0M 0 64.0M 0% /var/lib/docker/containers/0db3a0057c9194329fbacc4d5d94ab40eb2babe06dbb180f72ad96c8ff721632/shm
shm 64.0M 0 64.0M 0% /var/lib/docker/containers/6c17379244983233c7516062979684589c24b661bc203e6e1d53904dd7de167f/shm
tmpfs 3.9G 12.0K 3.9G 0% /opt/rke/var/lib/kubelet/pods/ea5b0e7d-18d6-11e9-86c9-fa163ebea4e5/volumes/kubernetes.io~secret/canal-token-gcxzd
tmpfs 3.9G 12.0K 3.9G 0% /opt/rke/var/lib/kubelet/pods/eab6dac4-18d6-11e9-86c9-fa163ebea4e5/volumes/kubernetes.io~secret/cattle-token-lbpxh
tmpfs 3.9G 8.0K 3.9G 0% /opt/rke/var/lib/kubelet/pods/eab6dac4-18d6-11e9-86c9-fa163ebea4e5/volumes/kubernetes.io~secret/cattle-credentials
tmpfs 3.9G 12.0K 3.9G 0% /opt/rke/var/lib/kubelet/pods/5c672b02-18df-11e9-a246-fa163ebea4e5/volumes/kubernetes.io~secret/nginx-ingress-serviceaccount-token-vc522
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/c29dc914ee801d2b36d4d2b688e5b060be6297665187f1001f9190fc9ace009d/merged
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/0591531eb89d598a8ef9bf49c6c21ea8250ad08489372d3ea5dbf561d44c9340/merged
shm 64.0M 0 64.0M 0% /var/lib/docker/containers/c89f839b36e0f7317c78d806a1ffb24d43a21c472a2e8a734785528c22cce85b/shm
shm 64.0M 0 64.0M 0% /var/lib/docker/containers/33050b02fc38091003e6a18385446f48989c8f64f9a02c64e41a8072beea817c/shm
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/a81da21f41c5c9eb2fb54ccdc187a26d5899f35933b4b701139d30f1af3860a4/merged
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/f6d546b54d59a29526e4a9187fb75c22c194d28926fca5c9839412933c53ee9d/merged
shm 64.0M 0 64.0M 0% /var/lib/docker/containers/7b0f9471bc66513589e79cc733ed6d69d897270902ffba5c9747b668d0f43472/shm
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/cae4765e9eb9004e1372b4b202e03a2a8d3880c918dbc27c676203eef7336080/merged
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/81ee00944f4eb367d4dd06664a7435634916be55c1aa0329509f7a277a522909/merged
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/7888843c2e76b5c3c342a765517ec06edd92b9eab25d26655b0f5812742aa790/merged
tmpfs 3.9G 12.0K 3.9G 0% /opt/rke/var/lib/kubelet/pods/c19a2ca3-18df-11e9-a246-fa163ebea4e5/volumes/kubernetes.io~secret/default-token-nzc2d
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/4d1c7efa3af94c1bea63021b594704db4504d4d97f5c858bdb6fe697bdefff9b/merged
shm 64.0M 0 64.0M 0% /var/lib/docker/containers/e10b7da6d372d241bebcf838e2cf9e6d86ce29801a297a4e7278c7b7329e895d/shm
overlay 7.4G 5.2G 1.8G 74% /var/lib/docker/overlay2/50df5234e85a2854b27aa8c7a8e483ca755803bc8bf61c25060a6c14b50a932c/merged
I already tried to prune all docker systems on all nodes and reconfigured and restarted all.
Is it may be connected with the fact that all the volumes have a limit of 7.4 GB?
How can I increase the ephemeral-storage therefore?
Thanks a lot
Is it may be connected with the fact that all the volumes have a limit of 7.4 GB?
You really have a single volume /dev/vda1
and multiple mount points and not several volumes with 7.4GB
Not sure where you are running Kubernetes but that looks like a virtual volume (in a VM). You can increase the size in the VM configuration or cloud provider and then run this to increase the size of the filesystem:
ext4:
$ resize2fs /dev/vda1
xfs:
$ xfs_growfs /dev/vda1
Other filesystems will have their own commands too.
The most common issue for running out of disk space on the master(s) is log files, so if that's the case you can set up a cleanup job for them or change the log size configs.