mwang1026
commented
3 years ago BACKUP
- https://github.com/cockroachdb/docs/issues/10634 -- Explain backup directory structure
- https://github.com/cockroachdb/docs/issues/10495 and https://github.com/cockroachdb/docs/issues/9809 -- how the nodes interact with BACKUP (and why every node needs to be able to talk to the storage bucket)
- A pointed to the gc.ttl text that we've now created (related https://github.com/cockroachdb/docs/issues/9940 but I think this could be a bigger task including backlinks from the gc.ttl zone config page, where we probably want a warning there about impact to backups, changefeeds)
- https://github.com/cockroachdb/docs/issues/7204 -- I think the relevant notion in this is BACKUP integrity -- what and how do we checksum? What is our encryption recommendation (e.g. use storage provider's on disk encryption?). related: https://github.com/cockroachdb/docs/issues/4700
RESTORE
- https://github.com/cockroachdb/docs/issues/10516 -- restoring does not restore MVCC history; it's a snapshot as of the system time you restored to
- I think we have https://github.com/cockroachdb/docs/issues/8433 better covered, but if not I think more about the various syntaxes you can use to restore from a collection of backups -- relates to the first item in the BACKUP section above
- https://github.com/cockroachdb/docs/issues/8541 is a bit outdated as we've made improvements since then
- https://github.com/cockroachdb/docs/issues/8024 -- this one is interesting, insofar as we don't prevent a user from creating stuff while restore is running a full cluster restore -- we might want to warn people off of that in general
- https://github.com/cockroachdb/docs/issues/8225 hits on a general theme of "what do we restore when we do a cluster restore". I could help enumerate this
Kathryn Hancox (kathancox) commented:
Our plan is to start compiling questions as an FAQ section around BACKUP and RESTORE.
@mwang1026 We can start adding related issues / comments to this issue for scope.
Jira Issue: DOC-1578