MDEV-36760 log_t::append_prepare_wait(): Bogus assertion on write_lsn #4038
+7
−8
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Outdated suggestions cannot be applied.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.
Description
log_t::append_prepare_wait()
: Do not attempt to readlog_sys.write_lsn
because it is not protected bylog_sys.latch
but bywrite_lock
, which we cannot hold here. The assertion could fail iflog_t::write_buf()
is executing concurrently, and it has not yet executedlog_write_buf()
or updatedlog_sys.write_lsn
.Fixes up #3925
Release Notes
N/A. This replaces a bogus debug assertion.
How can this PR be tested?
Run a write-heavy workload with the default
innodb_log_buffer_size
and possiblyinnodb_flush_log_at_trx_commit=0
. Possibly, add a sleep tolog_write_buf()
. If running on Linux/dev/shm
, be sure to compile withcmake -DWITH_INNODB_PMEM=OFF
in order to cover the correct code path.I tested this by running the default
mtr
suites whilemysql-test/var
was a symbolic link to a SATA device, as well as when it was a regular directory on/dev/shm
. That covered both code paths that are affected by these changes.Basing the PR against the correct MariaDB version
main
branch.PR quality check