Details
-
Bug
-
Resolution: Fixed
-
Critical
-
Cheshire-Cat
-
Untriaged
-
1
-
Unknown
Description
Can be easily reproduced by introducing a crash in perform_actual_join() like so:
$ git diff
|
diff --git i/src/ns_cluster.erl w/src/ns_cluster.erl
|
index a72c95aab..1387d3c3f 100644
|
--- i/src/ns_cluster.erl
|
+++ w/src/ns_cluster.erl
|
@@ -1327,6 +1327,8 @@ perform_actual_join(RemoteNode, NewCookie, ChronicleInfo) ->
|
ns_cluster_membership:prepare_to_join(RemoteNode, NewCookie),
|
ok = chronicle_local:prepare_join(ChronicleInfo),
|
|
+ exit(crash),
|
+
|
%% reload is needed to reinitialize ns_config's cache after
|
%% config cleanup ('erase' causes the problem, but it looks like
|
%% it's not worth it to add proper 'erase' support to ns_config)
|
Also seen in MB-46040: https://cb-jira.s3.us-east-2.amazonaws.com/logs/MB46040/output_file.zip
Attachments
Issue Links
- backports to
-
MB-49735 [Backport MB-46153] Node that fails to join the cluster might destroy all the data in that cluster
-
- Closed
-