Consider supporting other data stores #510
Replies: 2 comments 1 reply
-
Sounds good to me, I've been toying with using sqlite also for ease of dev/test. sqlite is easier as its still SQL, but getting the storage interface correct would be easier if we have a kv store in mind too. |
Beta Was this translation helpful? Give feedback.
-
Long time lurker here; I think flexibility in back-end datastores would be awesome. On the thread of using KV stores ala consul; would there be any advantage to integrating something like |
Beta Was this translation helpful? Give feedback.
-
Tinkerbell relies on Postgres as its datastore for hardware, template, and workflow data (and a few other workflow tables). Postgres is a fantastic database, but managing high-availability with graceful failover is not trivial.
Using an alternative data store that better supports failure would better help operators to have higher availability and not require downtime for upgrades or failover. There are several great open source options like etcd or consul that could be discussed in a later proposal. Both projects support watches that could support streaming workflow data to workers or other clients
I realize this is not a small change, and this isn't a formal technical proposal, but just a tracking discussion for the concept.
Beta Was this translation helpful? Give feedback.
All reactions