-
Bug
-
Resolution: Fixed
-
High
-
None
-
13
-
Severity 3 - Minor
-
Issue Summary
When performing a search after a page title, we have noticed that there are two links returned, one is the correct page link, but there is also a second page link which returns a Page Not found error. This issue only happened during migration from server to cloud, and only when the users migrated their site/space twice. The first import is a test run, so users deleted the spaces for the first import, they imported the space again, and the search results now contain deleted page links from the first import.
Environment
Atlassian Cloud
Steps to Reproduce
- Import a Confluence space with multiple pages as a test
- Delete the space
- Import again the space
- Search after a page title
- Notice that two links are displayed
Expected Results
Search should return only relevant page links, not the deleted page links
Actual Results
The search results contains also cache deleted pages from previous imports.
Notes
We have noticed that even if we run a Content indexing and Clear cache for the site, the issue still persists.
Workaround
Please contact support to open a request for CSRE. They have permission to run Elasticsearch query and update the sequence in the database
- Run an Elasticsearch query to find out the max version in the index. Send POST query to [http://|http:]{confluence_hostname}:9200/csp_{confluence_shard_name}_confluence_2/_search. Body:
{ "size":0, "query":{ "term": { "tenantId": $tenant_Id } }, "aggs" : { "max_version" : { "max" : { "field" : "_version" } } } }
- Bump the document_version_seq to max version in the index + 1. For example, if the max _version field in the index is 4301, then we will need to update the document_version_seq to version 4302
SELECT last_value FROM document_version_seq;
SELECT setval('document_version_seq', 111, true);
- Run re-index again.
Reproduction Steps for Local Test Environment
Root Cause:
1.
|
[Follow up] Finding root-cause for CONFCLOUD-66503 |
|
Closed | Unassigned |