[MXS-3455] Got packets out of order Created: 2021-03-24  Updated: 2021-08-23  Resolved: 2021-08-02

Status: Closed
Project: MariaDB MaxScale
Component/s: Core
Affects Version/s: 2.5.5
Fix Version/s: 2.5.14

Type: Bug Priority: Major
Reporter: Alex Assignee: Unassigned
Resolution: Fixed Votes: 0
Labels: None
Environment:

Debian



 Description   

We have two servers behind maxscale. Master is running on InnoDB, Slave is running on TokuDB. On slave we see these kind a messages in slow log and error log.

"Change user" command lasts for a second and "Got packets out of order" in error.log about same connection.

Mysql slow log:

# 
# # Time: 700101  3:00:01
# User@Host: meshok7[meshok7] @ sensored [10.0.0.1]
# Thread_id: 17142520  Schema: meshok  QC_hit: No
# Query_time: 1.002349  Lock_time: 0.000000  Rows_sent: 0  Rows_examined: 0
# Rows_affected: 0  Bytes_sent: 18446744073709546879
SET timestamp=1;
# administrator command: Change user;

error.log:

2021-03-24 10:32:01 17142520 [Warning] Aborted connection 17142520 to db: 'meshok' user: 'meshok7' host: 'sensored' (Got packets out of order)

show global variables like '%version%';
+-----------------------------------+------------------------------------------------------------+
| Variable_name                     | Value                                                      |
+-----------------------------------+------------------------------------------------------------+
| in_predicate_conversion_threshold | 1000                                                       |
| protocol_version                  | 10                                                         |
| slave_type_conversions            |                                                            |
| system_versioning_alter_history   | ERROR                                                      |
| system_versioning_asof            | DEFAULT                                                    |
| tls_version                       | TLSv1.1,TLSv1.2,TLSv1.3                                    |
| tokudb_version                    | 5.6.49-89.0                                                |
| version                           | 10.4.17-MariaDB-1:10.4.17+maria~stretch-log                |
| version_comment                   | mariadb.org binary distribution                            |
| version_compile_machine           | x86_64                                                     |
| version_compile_os                | debian-linux-gnu                                           |
| version_malloc_library            | jemalloc 3.6.0-0-g46c0af68bd248b04df75e4f92d5fb804c3d75340 |
| version_source_revision           | 5fbfdae130950d0a5a07d4b909f3bf1ff0498d34                   |
| version_ssl_library               | OpenSSL 1.1.0l  10 Sep 2019                                |
| wsrep_patch_version               | wsrep_26.22                                                |
+-----------------------------------+------------------------------------------------------------+



 Comments   
Comment by markus makela [ 2021-03-24 ]

Can you reproduce this with the latest release?

Comment by Alex [ 2021-03-24 ]

Will try it now.

Another problem is that casual_reads = global does not work as expected.

sometimes we insert row in DB
then immediately reading it by primary key and getting zero rows in response.

Comment by Alex [ 2021-03-24 ]

Just installed 2.5.9.

'Got packets out of order' persists

Comment by markus makela [ 2021-03-25 ]

My apologies, I thought that this was a duplicate of MXS-3436 and that the fix for it was in the 2.5.9 release but it seems to be in the upcoming 2.5.10 release. Once that release is out, please try to reproduce it again.

For the causal_reads=global problem, could you open a separate issue for that?

Comment by markus makela [ 2021-07-12 ]

alexcool can you test this with the latest 2.5 release?

Comment by markus makela [ 2021-08-02 ]

I'll close this as Fixed in 2.5.14 as I think this might be related to other fixes done in it. If you can, please confirm whether this is still a problem and if it is, we'll reopen this.

Comment by Alex [ 2021-08-23 ]

Looks fine on 2.5.15

Generated at Thu Feb 08 04:21:33 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.