Comments on: How to Read Simplified SHOW REPLICA STATUS Output https://www.percona.com/blog/how-to-read-simplified-show-replica-status/ Wed, 27 Sep 2023 14:21:56 +0000 hourly 1 https://wordpress.org/?v=6.5.2 By: Evgeny Gelfand https://www.percona.com/blog/how-to-read-simplified-show-replica-status/#comment-10974177 Wed, 27 Sep 2023 14:21:56 +0000 https://www.percona.com/blog/?p=90914#comment-10974177 Thank you for your response.
I apologize for the misformulation of my question.
Let me clarify:
Imagine a scenario where we have group replication that accepts changes from another data center:
Site A → Replication → Site B
On Site A, there is a source database running without GTID enabled, referred to as SDB.
On Site B, there is a Group Replication cluster (GRC) consisting of three servers: M1 (master), S1 (slave), and S2 (slave).
The replication channel, configured with ASSIGN_GTIDS_TO_ANONYMOUS_TRANSACTIONS, is defined on M1.
However, when M1 experiences an outage, a new master is automatically chosen.
My objective is to relocate the replication channel from the currently down M1 to the new master. I aim to ensure uninterrupted replication from SDB, regardless of the current master within the GRC.
While I can identify the newly chosen master using ProxySQL, the challenge lies in the channel metadata not being replicated. There are relevant tables under the mysql database, including:

  • slave_master_info
  • slave_relay_log_info
  • slave_worker_info

It appears that tables like mysql.slave_XXX are not replicated and do not get updated on the replicas within the group replication cluster. Interestingly, other tables, such as mysql.user, are successfully replicated.
This situation raises questions about whether this behavior is by design or possibly indicative of a bug.
Thank you,
Evgeny

]]>
By: Kedar Vaijanapurkar https://www.percona.com/blog/how-to-read-simplified-show-replica-status/#comment-10974175 Wed, 27 Sep 2023 07:36:25 +0000 https://www.percona.com/blog/?p=90914#comment-10974175 Hi Evgeny,

The table, mysql.slave_worker_info, doesn’t seem to be related to group replication, but standard replication. You can confirm the group replication status by querying

I encourage you to search MySQL group replication articles in Percona blog (there are plenty) or even head over to the Percona Forum to get solution from experts.

Thanks,
K

]]>
By: Evgeny Gelfand https://www.percona.com/blog/how-to-read-simplified-show-replica-status/#comment-10974169 Sat, 23 Sep 2023 02:10:55 +0000 https://www.percona.com/blog/?p=90914#comment-10974169 Hello.
Perhaps I am running group replication and the master is going down, how I can ensure that the new master will start from the place old master is left?
the table: mysq.slave_worker_info seems not replicated to the slaves in group replication.
What is possibilities please?

Thank you very much
Evgeny

]]>