mirror of
https://github.com/fleetdm/fleet
synced 2026-05-24 09:28:54 +00:00
8d18e67ab6
For #20046, I found this while trying to get #20085 merged, as MySQL 8 tests were consistenly failing (5 times in a row) with this when integration tests are run in parallel. I used the `lhotari/action-upterm@v1` action to hook an SSH session into the worker, and I noticed different errors in different runs, but all on the same lines of this sample: ``` mysql> SELECT * FROM replication_applier_status_by_worker LIMIT 1 OFFSET 0\G *************************** 1. row *************************** CHANNEL_NAME: WORKER_ID: 1 THREAD_ID: NULL SERVICE_STATE: OFF LAST_ERROR_NUMBER: 1146 LAST_ERROR_MESSAGE: Worker 1 failed executing transaction 'ANONYMOUS' at master log bin.000003, end_log_pos 25540153; Error 'Table 'server_datastore_mysql_TestTeams.host_mdm_apple_bootstrap_packages' doesn't exist' on query. Default database: 'server_datastore_mysql_TestTeams'. Query: 'TRUNCATE TABLE host_mdm_apple_bootstrap_packages' LAST_ERROR_TIMEST ``` This uses [REPLICATE_DO_DB](https://dev.mysql.com/doc/refman/5.7/en/change-replication-filter.html) to configure the replica at runtime to replicate only the databases that are configured with replicas and prevent any unexpected behavior. # Checklist for submitter If some of the following don't apply, delete the relevant line. <!-- Note that API documentation changes are now addressed by the product design team. --> - [x] Added/updated tests - [x] Manual QA for all new/changed functionality |
||
|---|---|---|
| .. | ||
| cached_mysql | ||
| filesystem | ||
| mysql | ||
| mysqlredis | ||
| redis | ||
| s3 | ||