Details
-
Bug
-
Resolution: Fixed
-
Critical
-
4.6.3, 5.0.0
-
Triaged
-
Unknown
Description
./testrunner -i INI_FILE.ini -t ent_backup_restore.enterprise_backup_merge_tests.EnterpriseBackupMergeTest.test_multiple_backups_merges,actions=backup_with_ops:3,failover_with_ops:graceful,backup_with_ops:2,recover_with_ops:full,merge:1&5,backup_with_ops:3,failover_with_ops:graceful,backup_with_ops:2,recover_with_ops:full,merge:2&6,merge:1&2,nodes_init=3,skip_cleanup=True
This testcase fails consistently at the end when we restore from the merged backup and validate the data - steps followed:
1. Setup cluster host and load up data
2. Backup 3 times while CRUD ops are on in parallel on the host
3. Failover a node gracefully while CRUD ops are on in parallel on the host
4. Backup 2 times while CRUD ops are on in parallel on the host
5. Recover failed over node with full recovery while CRUD ops are on in parallel on the host
6. Merge all 5 backups taken so far
7. Repeat 2 to 6
8. Merge 2 backups from the two iterations
9. Restore onto a cluster and validate data between source and destination cluster
Step 9 fails as there is a data mismatch between source and destination cluster - attaching backup log, zip of backup directory, console output and screenshot showing data mismatch in UI - mismatch is with the 'mutated' field always