feat(csi-driver): add pre-publish hook to cleanup stale entries #905
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
Abhinandan-Purkait
force-pushed
the
pre-publish-cp
branch
3 times, most recently
from
77f7bfb to
3fa1fe2
Compare
dsharma-dc
reviewed
Dec 12, 2024
dsharma-dc
approved these changes
Dec 12, 2024
Abhinandan-Purkait
force-pushed
the
pre-publish-cp
branch
2 times, most recently
from
f1d7cd4 to
18dc858
Compare
dsharma-dc
reviewed
Dec 12, 2024
dsharma-dc
reviewed
Dec 12, 2024
Abhinandan-Purkait
force-pushed
the
pre-publish-cp
branch
2 times, most recently
from
1d6ead3 to
3fedfd3
Compare
tiagolobocastro
approved these changes
Dec 12, 2024
Abhinandan-Purkait
force-pushed
the
pre-publish-cp
branch
3 times, most recently
from
273cea9 to
cd397c7
Compare
|
bors merge |
bors-openebs-mayastor bot
pushed a commit
that referenced
this pull request
Dec 12, 2024
905: feat(csi-driver): add pre-publish hook to cleanup stale entries r=Abhinandan-Purkait a=Abhinandan-Purkait Scenario: If an node undergoes a network partition and the application pods on it are moved by out-of-service taint then it can happen that the stale mount paths and the nvme controllers get left around, because CSI NodeUnstage was never called. Once the node comes back and joins the cluster and the application pod comes back to the same node, it can happen that the stale nvme path gets live again causing the existing buffer data to be flushed potentially causing data corruption. Changes: This adds a pre-publish hook that can intercept the ControllerPublish and issue cleanup on the node if stale entries are detected. Co-authored-by: Abhinandan Purkait <[email protected]>
|
Build failed: |
|
bors try |
tryBuild succeeded: |
Abhinandan-Purkait
force-pushed
the
pre-publish-cp
branch
from
499e665 to
827986c
Compare
|
bors merge |
bors-openebs-mayastor bot
pushed a commit
that referenced
this pull request
Dec 12, 2024
905: feat(csi-driver): add pre-publish hook to cleanup stale entries r=Abhinandan-Purkait a=Abhinandan-Purkait Scenario: If an node undergoes a network partition and the application pods on it are moved by out-of-service taint then it can happen that the stale mount paths and the nvme controllers get left around, because CSI NodeUnstage was never called. Once the node comes back and joins the cluster and the application pod comes back to the same node, it can happen that the stale nvme path gets live again causing the existing buffer data to be flushed potentially causing data corruption. Changes: This adds a pre-publish hook that can intercept the ControllerPublish and issue cleanup on the node if stale entries are detected. Co-authored-by: Abhinandan Purkait <[email protected]>
|
bors cancel |
|
Canceled. |
|
bors try |
tryBuild succeeded: |
Abhinandan-Purkait
force-pushed
the
pre-publish-cp
branch
from
93a2ebf to
ac33da7
Compare
Signed-off-by: Abhinandan Purkait <[email protected]>
Abhinandan-Purkait
force-pushed
the
pre-publish-cp
branch
from
ac33da7 to
a2f80e2
Compare
|
bors merge |
|
Build succeeded: |
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.
Scenario:
If an node undergoes a network partition and the application pods on it are moved by out-of-service taint then it can happen that the stale mount paths and the nvme controllers get left around, because CSI NodeUnstage was never called.
Once the node comes back and joins the cluster and the application pod comes back to the same node, it can happen that the stale nvme path gets live again causing the existing buffer data to be flushed potentially causing data corruption.
Changes:
This adds a pre-publish hook that can intercept the ControllerPublish and issue cleanup on the node if stale entries are detected.