[Flaky test] data_node_bootstrap fails with connection error

[mkindahl]

Which test is flaky?

data_node_bootstrap

Since when has the test been flaky?

November 09, 2022

Link to the failed test run

https://github.com/timescale/timescaledb/actions/runs/3426036397/jobs/5707419134

Log output

diff -u /mnt/d/a/timescaledb/timescaledb/tsl/test/expected/data_node_bootstrap.out /mnt/d/a/timescaledb/timescaledb/build_wsl/tsl/test/results/data_node_bootstrap.out
--- /mnt/d/a/timescaledb/timescaledb/tsl/test/expected/data_node_bootstrap.out  2022-11-09 07:08:33.625798700 +0000
+++ /mnt/d/a/timescaledb/timescaledb/build_wsl/tsl/test/results/data_node_bootstrap.out 2022-11-09 07:25:21.745537700 +0000
@@ -405,202 +405,4 @@
 CREATE SCHEMA ts_non_default AUTHORIZATION :ROLE_CLUSTER_SUPERUSER;
 SET client_min_messages TO NOTICE;
 \c access_node :ROLE_CLUSTER_SUPERUSER
--- Add data node and delete it under error suppression. We want to
--- avoid later tests to have random failures because the add succeeds.
-\set ON_ERROR_STOP 0
-SELECT node_name, database, node_created, database_created, extension_created
-FROM ts_non_default.add_data_node(
   .
   .
   .
+\connect: connection to server at "localhost" (127.0.0.1), port 55432 failed: SSL SYSCALL error: EOF detected

Reason for flakiness

No response

[svenklemm]

multinode has been removed