We should both backup etcd data stores to bacula and prepare a script / procedure people can use to easily restore the old state in case of emergency.
Description
Description
Details
Details
Related Gerrit Patches:
| operations/puppet : production | etcd: perform backups to /srv/backups/etcd, bacula |
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Open | None | T135122 Reduce etcd technical debt | |||
| Open | None | T135128 Turn on etcd TLS for intra-cluster communications | |||
| Resolved | Joe | T135129 Create backup/restore scripts for etcd |
Event Timeline
Comment Actions
Change 294916 had a related patch set uploaded (by Giuseppe Lavagetto):
etcd: perform backups to /srv/backups/etcd, bacula
Comment Actions
Change 294916 merged by Giuseppe Lavagetto:
etcd: perform backups to /srv/backups/etcd, bacula
Comment Actions
I have created an "etcd recovery script generator", that can be run and it proved to work with the labs cluster. It can be found at P3855
It can be used to generate disaster recovery instruction by indicating the cluster to recover and the host where a valid backup is thought to be.
It will take the user through all the stages of having a running cluster.