Details
-
Bug
-
Resolution: Done
-
Critical
-
Cheshire-Cat
-
Untriaged
-
-
1
-
Unknown
Description
Build : 7.0.0-4266
Test : -test tests/fts/cheshire-cat/test_fts_steady_state_cheshire_cat_coll_crud.yml -scope tests/fts/cheshire-cat/scope_fts_cheshire_cat.yml
Scale : 2
Iteration : 1st
In the updated system test, apart from creating 20 FTS indexes on collections (single collection index and multi-collection index), running FTS queries on these indexes, the test also has a step to continuously create and drop scopes and collections on 2 buckets - bucket4 & bucket5. There is another step to constantly create FTS indexes on any collection in these buckets. The drop of scope and collection would in turn drop the FTS index.
With this test, within 1 hour of running it, FTS service becomes unresponsive to any HTTP calls. Cannot even list indexes on any FTS nodes in the cluster.
HQ-Computer:~ mihirkamdar$ curl -v -u Administrator:password 172.23.121.165:8094/api/index
|
* Trying 172.23.121.165...
|
* TCP_NODELAY set
|
* Connected to 172.23.121.165 (172.23.121.165) port 8094 (#0)
|
* Server auth using Basic with user 'Administrator'
|
> GET /api/index HTTP/1.1
|
> Host: 172.23.121.165:8094
|
> Authorization: Basic QWRtaW5pc3RyYXRvcjpwYXNzd29yZA==
|
> User-Agent: curl/7.54.0
|
> Accept: */*
|
>
|
* Empty reply from server
|
* Connection #0 to host 172.23.121.165 left intact
|
*
|
Also seeing MB-43490 on all FTS nodes. The test cannot proceed further.
KV nodes in the cluster : 172.23.104.16, 172.23.104.17, 172.23.104.18, 172.23.104.19, 172.23.104.21, 172.23.104.23
FTS nodes in the cluster : 172.23.121.165, 172.23.96.30, 172.23.97.77, 172.23.97.82, 172.23.97.83