Closed danielealbano closed 1 year ago
Patch coverage: 100.00
% and project coverage change: -0.03
:warning:
Comparison is base (
3eb32e4
) 76.83% compared to head (97ace96
) 76.81%.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.
This PR fixes a sneaky issue in the hashtable locking, the construct used to get the worker index only once per thread was actually being called only the very first time by only one thread causing all the other threads to have a worker_index of 0.
As consequence the locking was potentially generating overlappy transaction ids leading to a potential clash.
This behaviour was put in evidence by redis-benchmark which uses only 1 single key for all the requests unless the
--threads
option is passed.The PR also changes 2 other aspects: