How to Cleanup Incorrectly Configured Docker Storage
Problem
Docker cannot start up on the nodes due to incorrectly configured Storage. The following error is observed in docker logs journalctl -u docker
S-NOR-KUB02 dockerd[14104]: failed to start daemon: error initializing graphdriver: devmapper: Unable to take ownership of thin-pool (docker--vg-thinpool) that already has used data blocksIf storage for docker was manually configured without the required parameters the Platform9 implementation will not be able to utilize it.
Environment
- Platform9 Managed Kubernetes - All Versions
- Docker
- LVM
Procedure
This will delete all the data currently present in the Volume Group. If you wish to retain this data please take backup before this process.
1 - Make sure you have unmounted all the PV which are a part of this Volume Group. Once that is done you can remove the Logical volumes associated with them.
$ sudo lvremove -f <__vg_name__>/<__lv_name__>2 - To remove the Volume Group we have deactivate it first.
$ vgchange -an <__vg_name__>3 - Remove the VG
$ sudo vgremove <__vg_name__>4 - Once the Volume Group is deleted, follow the steps given in this article to setup Docker Storage for your Platform9 Managed Node