[storage] Add a progress shard for sinks and skip the snapshot when possible #31152
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
bkirwi
force-pushed
the
sink-progress
branch
6 times, most recently
from
e2cec57 to
aa0797f
Compare
bkirwi
force-pushed
the
sink-progress
branch
8 times, most recently
from
0b12035 to
b24b2aa
Compare
bkirwi
changed the title
|
Still doing some testing, but I think this feels more or less right. @aljoscha or @petrosagg, able to take a look? |
bkirwi
commented
Feb 6, 2025
|
|
||
| > SELECT status | ||
| FROM mz_internal.mz_sink_statuses | ||
| WHERE name = 'snk' | ||
| dropped |
There was a problem hiding this comment.
Previously, the read frontier was immediately set to {} for sinks, so having the write frontier progress to {} would mark the sink dropped. After this PR, the read frontier can evolve independently, so the sink is not immediately removed from introspection.
Looking at https://github.com/MaterializeInc/database-issues/issues/8842 - this seems like an improvement, though there may be more to do. (In the future, not on this PR.)
bkirwi
force-pushed
the
sink-progress
branch
2 times, most recently
from
1941c36 to
e008647
Compare
Tweaking tests slightly to assert properties that are correct in both the old and new implementations.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Motivation
Design sketch is here: https://github.com/MaterializeInc/database-issues/issues/8603#issuecomment-2611104841
Tips for reviewer
Adding a collection for sinks makes them much more like other types of things that the storage controller manages, which allows (and forces) us to handle them as a type of collection instead of a separate thing. Reworking this is much of the size of the diff - though the net result is slightly less / more uniform controller code, which is nice.
I've left a number of
// TODO(sinks)todos for possible followups that I didn't want to bite off here.Checklist
$T ⇔ Proto$Tmapping (possibly in a backwards-incompatible way), then it is tagged with aT-protolabel.