crimson: take obc lock during push commit on primary#61561
Merged
crimson: take obc lock during push commit on primary#61561
Conversation
Should make it a bit easier to find the two kinds of message handlers. Signed-off-by: Samuel Just <sjust@redhat.com>
…ome formatting changes Signed-off-by: Samuel Just <sjust@redhat.com>
…ing changes Signed-off-by: Samuel Just <sjust@redhat.com>
…coroutine Signed-off-by: Samuel Just <sjust@redhat.com>
…coroutine Signed-off-by: Samuel Just <sjust@redhat.com>
…e to coroutine Signed-off-by: Samuel Just <sjust@redhat.com>
… to coroutine Signed-off-by: Samuel Just <sjust@redhat.com>
…object if complete Signed-off-by: Samuel Just <sjust@redhat.com>
…le_pull_response Signed-off-by: Samuel Just <sjust@redhat.com>
Avoids extra lookup. Signed-off-by: Samuel Just <sjust@redhat.com>
Signed-off-by: Samuel Just <sjust@redhat.com>
Already checked in handle_recovery_op. Signed-off-by: Samuel Just <sjust@redhat.com>
1a68b82 to
4e70fbb
Compare
Contributor
Author
|
jenkins test make check |
Matan-B
approved these changes
Jan 30, 2025
Comment on lines
-1033
to
-1036
| if (pg.can_discard_replica_op(*m)) { | ||
| DEBUGDPP("discarding {}", pg, *m); | ||
| return seastar::now(); | ||
| } |
Contributor
There was a problem hiding this comment.
The same can be applied for pull/pull_response, right?
Contributor
Author
There was a problem hiding this comment.
Good call, but it can be a future PR
Comment on lines
+1348
to
+1356
| if (pg.is_primary()) { | ||
| return handle_pull_response( | ||
| boost::static_pointer_cast<MOSDPGPush>(m)); | ||
| } else { | ||
| return handle_push(boost::static_pointer_cast<MOSDPGPush>(m)); | ||
| } |
Contributor
There was a problem hiding this comment.
nit:
crimson/.../replicated_recovery_backend: route pull_response earlier
Let ReplicatedRecoveryBackend::handle_recovery_op route
pull_responses instead of ReplicatedRecoveryBackend::handle_push.
| pull_info.stat.num_objects_recovered++; | ||
| auto manager = pg.obc_loader.get_obc_manager( | ||
| recovery_waiter.obc); | ||
| manager.lock_excl_sync(); /* cannot already be locked */ |
Contributor
There was a problem hiding this comment.
This obc cannot already be locked since we've just recovered it?
Contributor
Author
There was a problem hiding this comment.
Yeah, we just recovered it so it can't be locked.
| seastar::make_ready_future<>()); | ||
| if (pull_info.recovery_progress.first) { | ||
| prepare_waiter = pg.obc_loader.with_obc<RWState::RWNONE>( | ||
| auto fut = pg.obc_loader.with_obc<RWState::RWNONE>( |
Contributor
There was a problem hiding this comment.
Should we move this bit to the manager as well? (creation of the WaitForObjectRecovery::obc that will be the one taking excl lock later on)
auto obc_orderer = pg.obc_loader.get_obc_orderer(pull_info.recovery_info.soid);
auto obc_manager = pg.obc_loader.get_obc_manager(
obc_orderer,
pull_info.recovery_info.soid);
co_await pg.obc_loader.load_and_lock(
obc_manager, RWState::RWNONE
).handle_error_interruptible(
crimson::ct_error::assert_all("unexpected error")
);
// TODO: move to ObjectContextLoader once constructing obc from attrset is supported
obc_manager.get_obc()->obs.oi.decode_no_oid(push_op.attrset.at(OI_ATTR),
....
Contributor
There was a problem hiding this comment.
No lock has to be taken here since we are not submitting the transaction yet and its the phase where we only create this obc?
Contributor
Author
There was a problem hiding this comment.
Good call, but it can be a future PR
Contributor
|
@athanatos, since this was tested I'm ok with merging as is and follow up on some of the comments later on. |
14 tasks
Let ReplicatedRecoveryBackend::handle_recovery_op route pushes between handle_push and handle_pull_response instead of ReplicatedRecoveryBackend::handle_push. Signed-off-by: Samuel Just <sjust@redhat.com>
…commit Fixes: https://tracker.ceph.com/issues/69412 Signed-off-by: Samuel Just <sjust@redhat.com>
4e70fbb to
6093f91
Compare
Contributor
Author
|
jenkins test make check |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
See https://tracker.ceph.com/issues/69412
Follow up PR will address replica side.
Test run looks good, failures listed below:
https://pulpito.ceph.com/sjust-2025-01-28_06:26:29-crimson-rados-wip-sjust-crimson-testing-2025-01-27-distro-default-smithi/
Show available Jenkins commands
jenkins retest this pleasejenkins test classic perfjenkins test crimson perfjenkins test signedjenkins test make checkjenkins test make check arm64jenkins test submodulesjenkins test dashboardjenkins test dashboard cephadmjenkins test apijenkins test docsjenkins render docsjenkins test ceph-volume alljenkins test ceph-volume toxjenkins test windowsjenkins test rook e2e