Category Archives: Monitoring

Extending replication instrumentation: an insight on transaction retries

MySQL 8.0.13 improves replication lag monitoring by extending the instrumentation for transaction transient errors. These temporary errors, which include lock timeouts caused by client transactions executing concurrently as the slave is replicating, do not stop the applier thread: instead, they cause a transaction to retry.…

More P_S Instrumentation for Group Replication

In order to improve observability and manageability of MySQL Group Replication (GR), we enhanced its instrumentation in MySQL 8.0.4. With these goals in mind, we have extended the usage of Performance Schema (P_S) from this plugin to the included Group Communication System (GCS) module as well as its submodule eXtended Communications (XCom) which are responsible for all communication involved in the dissemination of transactions to other MySQL instances.…

“Look: I/O thread is waiting for disk space!”

MySQL 8.0.1 introduced a work with replication threads mutexes in order to improve performance. In MySQL 8.0.2 the same work was extended, focusing in usability, and revamped how replication deals with disk-full conditions, improving the responsiveness of both monitoring commands and administrative commands such as KILL, as well as making status messages much more precise and helpful.…

Group Replication – Extending Group Replication performance_schema tables

In MySQL 8.0.2, users will see the additional columns in the existing Group Replication Performance Schema tables which will provide extended information about Group Replication. Now user can view role and MySQL version of each member of the group, which earlier required a complex set of query.…

MySQL 8.0.1 Replication Performance Improvements

MySQL 8.0.1 introduces new features and improvements to replication. The next sections summarize some related to performance.

1. WL#9556 – Writeset-based dependency tracking on the master

Transaction writesets can now be used to evaluate which transactions are independent and can be executed in parallel on the binary log applier on the asynchronous slaves.…