Reputation: 1209
I follows the steps on this page
STOP SLAVE;
SET GTID_NEXT="[THE GTID SET]";
BEGIN; COMMIT;
SET GTID_NEXT="AUTOMATIC";
START SLAVE;
to restore the slave. But in my case, the gtid_set are as
Retrieved_Gtid_Set: 8b6d4795-5ad3-11e6-a31f-00259077c77a:2369-2377
Executed_Gtid_Set: 8b6d4795-5ad3-11e6-a31f-00259077c77a:1-2372:2374,8be5b0ba-5ad3-11e6-a31f-0cc47a50d072:1-12
When I tried to inject empty transactions to the slave and restart the slave, the 'slave_SQL_Running' is still 'No'.
STOP SLAVE;
SET GTID_NEXT="8b6d4795-5ad3-11e6-a31f-00259077c77a:2377";
BEGIN; COMMIT;
SET GTID_NEXT=AUTOMATIC;
START SLAVE;
And it becomes
Retrieved_Gtid_Set: 8b6d4795-5ad3-11e6-a31f-00259077c77a:2369-2377
Executed_Gtid_Set: 8b6d4795-5ad3-11e6-a31f-00259077c77a:1-2372:2374:2377,8be5b0ba-5ad3-11e6-a31f-0cc47a50d072:1-12
and when new data is inserted to master, the slave still can't sync to the master. The status becomes:
Retrieved_Gtid_Set: 8b6d4795-5ad3-11e6-a31f-00259077c77a:2369-2378
Executed_Gtid_Set: 8b6d4795-5ad3-11e6-a31f-00259077c77a:1-2372:2374:2377,8be5b0ba-5ad3-11e6-a31f-0cc47a50d072:1-12
How can I make this work?
I don't want to do a full dump since there's lots of data, while fulldump takes a lot of time.
Upvotes: 0
Views: 2164
Reputation: 16
Please check below blog post.
https://www.abhinavbit.com/2019/05/gtid-replication-skip-transaction-using.html
In your case you should have run below command to skip the sql threads error.
STOP SLAVE;
SET GTID_NEXT="8b6d4795-5ad3-11e6-a31f-00259077c77a:2375";
BEGIN; COMMIT;
SET GTID_NEXT=AUTOMATIC;
START SLAVE;
Upvotes: 0