I just hit a problem with updatedb on k8s nodes. It stucks inside mounted external disks such as rbd and prevent them to unmount and thus migrate. Here is lsof tail
lsof: no pwd entry for UID 1000
updatedb. 743930 root cwd DIR 251,0 4096 12 /var/lib/kubelet/plugins/kubernetes.io/rbd/mounts/kube-image-kubernetes-dynamic-pvc-3c72e7ac-aece-11e8-b160-0a580af40542/alertmanager-db
updatedb. 743930 root 16r DIR 251,0 4096 2 /var/lib/kubelet/plugins/kubernetes.io/rbd/mounts/kube-image-kubernetes-dynamic-pvc-3c72e7ac-aece-11e8-b160-0a580af40542
updatedb. 743930 root 17r DIR 251,0 4096 12 /var/lib/kubelet/plugins/kubernetes.io/rbd/mounts/kube-image-kubernetes-dynamic-pvc-3c72e7ac-aece-11e8-b160-0a580af40542/alertmanager-db
IMHO it would be great to have an ability to specify custom post-install script to run on all servers. This may provide required flexibility and simplicity.
I just hit a problem with updatedb on k8s nodes. It stucks inside mounted external disks such as rbd and prevent them to unmount and thus migrate. Here is lsof tail
Here is processlist
It would be nice to have mlocate/updatedb disabled during node provisioning