max_allowed_packetsets an upper limit on the size of any single message between the MySQL server and clients, including replicas. If you are replicating large column values (such as might be found inTEXTorBLOBcolumns) andmax_allowed_packetis too small on the source, the source fails with an error, and the replica shuts down the replication I/O thread. Ifmax_allowed_packetis too small on the replica, this also causes the replica to stop the I/O thread.

Row-based replication currently sends all columns and column values for updated rows from the source to the replica, including values of columns that were not actually changed by the update. This means that, when you are replicating large column values using row-based replication, you must take care to setmax_allowed_packetlarge enough to accommodate the largest row in any table to be replicated, even if you are replicating updates only, or you are inserting only relatively small values.

On a multi-threaded replica (withreplica_parallel_workers > 0orslave_parallel_workers > 0), ensure that the system variablereplica_pending_jobs_size_maxorslave_pending_jobs_size_maxis set to a value equal to or greater than the setting for themax_allowed_packetsystem variable on the source. The default setting forreplica_pending_jobs_size_maxorslave_pending_jobs_size_max, 128M, is twice the default setting formax_allowed_packet, which is 64M.max_allowed_packetlimits the packet size that the source can send, but the addition of an event header can produce a binary log event exceeding this size. Also, in row-based replication, a single event can be significantly larger than themax_allowed_packetsize, because the value ofmax_allowed_packetonly limits each column of the table.

The replica actually accepts packets up to the limit set by itsreplica_max_allowed_packetorslave_max_allowed_packetsetting, which default to the maximum setting of 1GB, to prevent a replication failure due to a large packet. However, the value ofreplica_pending_jobs_size_maxorslave_pending_jobs_size_maxcontrols the memory that is made available on the replica to hold incoming packets. The specified memory is shared among all the replica worker queues.

The value ofreplica_pending_jobs_size_maxorslave_pending_jobs_size_maxis a soft limit, and if an unusually large event (consisting of one or multiple packets) exceeds this size, the transaction is held until all the replica workers have empty queues, and then processed. All subsequent transactions are held until the large transaction has been completed. So although unusual events larger thanreplica_pending_jobs_size_maxorslave_pending_jobs_size_maxcan be processed, the delay to clear the queues of all the replica workers and the wait to queue subsequent transactions can cause lag on the replica and decreased concurrency of the replica workers.replica_pending_jobs_size_maxorslave_pending_jobs_size_maxshould therefore be set high enough to accommodate most expected event sizes.