You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
…r RPW to rendezvous RPW (#7397)
This PR builds on #7386. I'll test both of them together on a racklette
(will put results in a comment on this PR).
This PR contributes to addressing (but doesn't fully address) a couple
issues:
* #7392: This PR adds useful output for the `crucible_dataset`
rendezvous table, but we still need useful output for `debug_dataset`
* #6999: This PR removes a step from execution that currently depends on
successful completion of earlier step to post datasets to sled-agents,
since we now confirm insertions are okay via inventory collection
I'm inclined to say this PR fully addresses a couple other issues, but
it's debatable:
* #7210: Are we okay with IP and port still being here now that this
table is just crucible datasets? (IMO it's still conceptually weird to
say a dataset has an IP/port, but maybe it's not as bad now?)
* #6998 `dataset` as a general table is now gone and replaced with more
specific rendezvous tables, so I think this one can be closed?
It would be nice if we could rely on inventory for "datasets we observe", and the blueprint for "datasets we want to have".
There exists a dataset table that's kinda redundant with the blueprint information, but exists for historical reasons:
omicron/schema/crdb/dbinit.sql
Lines 559 to 604 in 15b4c45
We used to have a similar table for services, but got rid of it in favor of blueprints.
It would be nice if we could do the same for datasets!
omicron.public.dataset
records to look up Crucible socket addressesomicron.public.dataset
records to track Crucible space usageThe text was updated successfully, but these errors were encountered: