I just did a run of the regression test where this test was the last one to finish by quite a lot. Key log entries:
[13:35:48.583](0.039s) # initializing database system by copying initdb template ... [13:35:52.397](0.108s) ok 5 - Check reset timestamp for 'test_tab1_sub' is newer after second reset. #### Begin standard error psql:<stdin>:1: NOTICE: created replication slot "test_tab2_sub" on publisher #### End standard error Waiting for replication conn test_tab2_sub's replay_lsn to pass 0/151E8C8 on publisher done [13:38:53.706](181.310s) ok 6 - Check that table 'test_tab2' now has 1 row. ... [13:38:54.344](0.294s) 1..13 I reran the test and it looks very different: [13:54:01.703](0.090s) ok 5 - Check reset timestamp for 'test_tab1_sub' is newer after second reset. ... Waiting for replication conn test_tab2_sub's replay_lsn to pass 0/151E900 on publisher ... [13:54:03.006](1.303s) ok 6 - Check that table 'test_tab2' now has 1 row. It looks to me like in the first run it took 3 minutes for the replay_lsn to catch up to the desired value, and in the second run, two seconds. I think I have seen previous instances where something similar happened, although in those cases I did not stop to record any details. Have others seen this? Is there something we can/should do about it? -- Robert Haas EDB: http://www.enterprisedb.com
