Setting up slave_parallel_type = LOGICAL_CLOCK dynamically not working in pxc

Description

Hi,

Here, not able to set the slave_parallel_type='LOGICAL_CLOCK' dynamically but able to change with cnf and restart:
node1 [localhost:26530] {msandbox} ((none)) > set global slave_parallel_type='LOGICAL_CLOCK';
ERROR 1105 (HY000): Percona-XtraDB-Cluster prohibits enabling slave_preserve_commit_order as it conflicts with galera multi-master commit order semantics
mysql [localhost:6739] {msandbox} ((none)) > show global variables like 'slave_%';
----------------------------------------------------------------------------------+

Variable_name

Value

----------------------------------------------------------------------------------+

slave_allow_batching

OFF

slave_checkpoint_group

512

slave_checkpoint_period

300

slave_compressed_protocol

OFF

slave_exec_mode

STRICT

slave_load_tmpdir

/home/vinodh.krishnaswamy/sandboxes/msb_5_7_29/tmp

slave_max_allowed_packet

1073741824

slave_net_timeout

60

slave_parallel_type

DATABASE

slave_parallel_workers

0

slave_pending_jobs_size_max

16777216

slave_preserve_commit_order

OFF

slave_rows_search_algorithms

TABLE_SCAN,INDEX_SCAN

slave_skip_errors

OFF

slave_sql_verify_checksum

ON

slave_transaction_retries

10

slave_type_conversions

 

----------------------------------------------------------------------------------+
17 rows in set (0.00 sec)
 

But when we edit the cnf like below and restarting makes the changes: (changed node2's cnf for testing)
relay_log_info_repository = TABLE
slave_parallel_type = LOGICAL_CLOCK
slave_parallel_workers = 4
relay_log_recovery = ON
 

after restart:
node2 [localhost:26531] {msandbox} ((none)) > show global variables like 'slave_paral%';
---------------------------------------+

Variable_name

Value

---------------------------------------+

slave_parallel_type

LOGICAL_CLOCK

slave_parallel_workers

4

---------------------------------------+
2 rows in set (0.01 sec)node2 [localhost:26531] {msandbox} ((none)) > show global variables like 'slave_p%';
--------------------------------------------+

Variable_name

Value

--------------------------------------------+

slave_parallel_type

LOGICAL_CLOCK

slave_parallel_workers

4

slave_pending_jobs_size_max

16777216

slave_preserve_commit_order

OFF

--------------------------------------------+
4 rows in set (0.01 sec)
 
node2 [localhost:26531] {msandbox} ((none)) > status
--------------
/home/vinodh.krishnaswamy/opt/mysql/pxc5.7.28/bin/mysql Ver 14.14 Distrib 5.7.28-31, for Linux (x86_64) using 6.2Connection id: 5
Current database:
Current user: msandbox@localhost
SSL: Not in use
Current pager: stdout
Using outfile: ''
Using delimiter: ;
Server version: 5.7.28-31-31.41-log Percona XtraDB Cluster binary (GPL) 5.7.28-31.41, Revision ef2fa88, wsrep_31.41
Protocol version: 10
Connection: Localhost via UNIX socket
Server characterset: latin1
Db characterset: latin1
Client characterset: utf8
Conn. characterset: utf8
UNIX socket: /tmp/mysql_sandbox26531.sock
Uptime: 13 min 26 secThreads: 4 Questions: 9 Slow queries: 0 Opens: 112 Flush tables: 1 Open tables: 27 Queries per second avg: 0.011
--------------
 

Please check why it is behaving like it.

 

Apart from this, noted from the doc, from 5.7.12rc, setting up slave_preserve_commit_order is not supported. But this is not related with slave_parallel_type's value setting issue above.

https://www.percona.com/doc/percona-xtradb-cluster/5.7/release-notes/Percona-XtraDB-Cluster-5.7.12-5rc1-26.16.html

  • Disabled switching of slave_preserve_commit_order to ON when running PXC in cluster mode, as it conflicts with existing multi-master commit ordering resolution algorithm in Galera.

 

 

Environment

None

AFFECTED CS IDs

CS0013494

Smart Checklist

Activity

Show:

Julia Vural March 4, 2025 at 9:27 PM

It appears that this issue is no longer being worked on, so we are closing it for housekeeping purposes. If you believe the issue still exists, please open a new ticket after confirming it's present in the latest release.

bartosz.gatz July 15, 2024 at 11:39 AM

Answering for Venki: these two issues are different.

bartosz.gatz June 24, 2024 at 11:50 AM

can you take a look at Kamil’s question? If this is a duplicate let’s close this one as a duplicate (and link it to the main bug). Thanks!

Michael Coburn November 27, 2021 at 11:14 PM

This would be a great feature for Services for migrations into PXC using async replication, in order to minimize replication lag thumbs up

Won't Do

Details

Assignee

Reporter

Labels

Priority

Smart Checklist

Created October 24, 2020 at 7:59 AM
Updated March 4, 2025 at 9:27 PM
Resolved March 4, 2025 at 9:27 PM

Flag notifications