K2V #293
No reviewers
Labels
No labels
action
check-aws
action
discussion-needed
action
for-external-contributors
action
for-newcomers
action
more-info-needed
action
need-funding
action
triage-required
kind
correctness
kind
ideas
kind
improvement
kind
performance
kind
testing
kind
usability
kind
wrong-behavior
prio
critical
prio
low
scope
admin-api
scope
background-healing
scope
build
scope
documentation
scope
k8s
scope
layout
scope
metadata
scope
ops
scope
rpc
scope
s3-api
scope
security
scope
telemetry
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference: Deuxfleurs/garage#293
Loading…
Reference in a new issue
No description provided.
Delete branch "k2v"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Specification:
Making a Gitea PR as this is easier to edit and to comment than on cryptpad
View spec at this URL
Implementation:
Testing:
Improvements:
experimental
feature flagWIP: K2V specificationto WIP: K2V26f44e5d85
to3d4d59e714
reverse
worksquick review of k2v spec. Overall looks good, see comments for details
@ -0,0 +13,4 @@
- K2V triplets exist separately from S3 objects. K2V triples don't exist for
the S3 API, and S3 objects don't exist for the K2V API.
- Values stored for triples have associated causality information, that enables
are triples and triplets the same thing? If they are different we should clarify how, and if they are the same, we should use only one word to name them.
@ -0,0 +172,4 @@
- **Internal server errors on updates do not mean that the update isn't stored.**
K2V will return an internal server error when it cannot reach a quorum of nodes on
which to save an updated value. However the value may still be stored on just one
node, which whill then propagate it to other nodes asynchronously via anti-entropy.
which will
@ -0,0 +653,4 @@
prefix: null,
start: null,
end: null,
singleItem: false,
maybe
singleItem
should be true by default for delete operation, that would be safer against faulty API usage (it would probably require being the same for SEARCH, for the sake of coherence)@ -0,0 +30,4 @@
- a partition key (`pk`), an utf8 string that defines in what partition the triple is
stored; triples in different partitions cannot be listed together, they must
be the object of different ReadItem or ReadBatch queries
ReadItem reads a single triplet, so I don't think it's affected? Also, are other *Batch affected? I assume no, but this should probably be explicited
e4443d720d
toeaca736a4f
WIP: K2Vto K2V