You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Found that tests may fail due to hang in seek_wait() loop on starting
and stopping instances. It happened, because instances were not synced
within default output timeout which is by default 120 seconds, while
replication sync happens only each 300 seconds by default. To fix it
replication_sync_timeout should be decreased to some value lower than
'no-output-timeout', so decided to set it to 100 seconds.
The issue looked like in tests:
--- replication/gh-5140-qsync-casc-rollback.result Mon Oct 19
17:29:46 2020
+++ /rw_bins/test/var/070_replication/gh-5140-qsync-casc-rollback.result
Mon Oct 19 17:31:35 2020
@@ -169,56 +169,3 @@
-- all the records are replayed one be one without yields for WAL writes, and
-- nothing should change.
test_run:cmd('restart server default')
- |
-test_run:cmd('restart server replica')
- | ---
- | - true
- | ...
-
Part of tarantool/tarantool#5504
0 commit comments