Fix unstable test: replication with parallel clients writing in different DBs (#11782)
Failure happens in FreeBSD daily: ``` *** [err]: Test replication with parallel clients writing in different DBs in tests/integration/replication-4.tcl Expected [::redis::redisHandle2 dbsize] > 0 (context: type eval line 19 cmd {assert {[$master dbsize] > 0}} proc ::test) ``` The test is failing because db 9 has no data (default db), and according to the log, we can see that db 9 does not have a key: ``` ### Starting test Test replication with parallel clients writing in different DBs in tests/integration/replication-4.tcl 3338:S 03 Feb 2023 00:15:18.723 - DB 11: 1 keys (0 volatile) in 4 slots HT. 3338:S 03 Feb 2023 00:15:18.723 - DB 12: 141 keys (0 volatile) in 256 slots HT. ``` We use `wait_for_condition` to ensure that parallel clients have written data before calling stop_bg_complex_data. At the same time, `wait_for_condition` is also used to remove the above `after 1000`, which can save time in most cases.
Please register or sign in to comment