Closed
Conversation
ioThreadsScaleUpStart() (which logs 'IO threads scale-up end') runs in the *next* beforeSleep() after CONFIG SET is processed. However, io_thread_1 may already flush the 'OK' reply to the socket before the main thread enters that next beforeSleep(), causing Tcl to call verify_log_message (which checks immediately, no waiting) before the log line is written. Under ASAN the main thread is slower due to memory instrumentation while the io thread's socket write is a simple syscall, widening the race window. Fix: add wait_for_condition for io_thread_scale_status == 'none' between the CONFIG SET and verify_log_message, matching the pattern already used in the 101-client case at line 173-177. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
…napshot Line 248 captures a snapshot via 'set info [r info threads]' and checks whether io_thread_2 exists in it. The original line 249 then issued a second 'r info threads' query inside the if-body. If io_thread_2 was torn down between those two RPCs (easy under ASAN where the main thread is slower), the second query returns '' for io_thread_2, causing assert_equal '' 0 to fail. Fix: reuse the already-captured $info for the clients assertion. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The 'Test child sending info' test fails under ASAN with:
COW info wasn't reported
Root cause: sendChildInfoGeneric() throttles /proc/self/smaps reads
(zmalloc_get_private_dirty) using CHILD_COW_DUTY_CYCLE=100. Under ASAN
the process virtual address space is roughly doubled by shadow memory, so
/proc/self/smaps is much larger and can take 2-10 seconds to parse. The
duty-cycle suppresses the NEXT measurement for cow_update_cost * 100
microseconds — potentially hundreds of seconds — making it impossible to
see a second incremental COW update within the original 8-second window.
Three targeted fixes for ASAN builds ($::asan == 1):
1. Increase wait_for_condition from 80*100ms to 300*100ms (30 s) so the
first slow smaps read can complete and the parent can absorb it via
run_with_period(1000).
2. Break after the first successful COW observation (after incr iteration
sets iteration=2), skipping the second iteration that would require a
second smaps read throttled away for minutes.
3. Relax 'assert_morethan_equal $iteration 2' to 'assert {$iteration >= 1}'
and change the final rdb_last_cow_size check to 'assert {$final_cow > 0}'
(instead of 90% of last active cow_size which can be 0 under ASAN when
bgsave completes before the first incremental report is visible).
Normal (non-ASAN) behaviour is unchanged.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
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
1 participant
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.
No description provided.