-
Type:
Improvement
-
Status: Done
-
Priority:
Medium
-
Resolution: Fixed
-
Affects Version/s: None
-
Fix Version/s: 2.0.5
-
Component/s: None
-
Labels:None
STR:
1) setup pxc
2) setup proxysql clustering for 3 instances
3) run `proxysql-admin --config-file=/etc/proxysql-admin.cnf --cluster-hostname=cluster1-pxc-0.cluster1-pxc.run-master-2303.svc.cluster.local --use-ssl=yes --enable --update-cluster --remove-all-servers --force` for all instances
4) run `proxysql-admin --config-file=/etc/proxysql-admin.cnf --cluster-hostname=cluster1-pxc-0.cluster1-pxc.run-master-2303.svc.cluster.local --use-ssl=yes --syncusers --add-query-rule`
5) add user 'user-test' and query rules for it
6) create new proxysql instance from the scratch, add it into proxysql clustering
7) run command 3) for newly created proxysql instance
8) run command 4) imidiatelly after the previous command
current result:
proxysql clustering don't have time to sync users to newly created instance from cluster configuration (between steps 7 and 8)
in result, the custom user and query rules for it (created in in step 5) become removed
expected result
`--syncusers` command shouldn't remove users for whole cluster
propoused solution:
1) make `syncusers` procedure should be part of `enable` stage
2) if `syncusers` procedure part of `enable` stage it shouldn't increase `version` and `epoch` in `runtime_checksums_values` table.
in such case user add `syncusers` into cron and custom users in existsing cluster never gona be cleaned during cluster scale up.