Bug #88607 mysql start group_replication Too many connections 2k+ then then mysql crash
Submitted: 22 Nov 2017 14:14 Modified: 7 Jan 2018 2:07
Reporter: yaodong feng Email Updates:
Status: No Feedback Impact on me:
None 
Category:MySQL Server: Group Replication Severity:S1 (Critical)
Version:mysql5.7.19 OS:CentOS (7)
Assigned to: MySQL Verification Team CPU Architecture:Any
Tags: mgr, MySql Crash

[22 Nov 2017 14:14] yaodong feng 
Description:
in my.cnf max_connections = 6000
show processlist connnections about 2300+

My operation is to change the master-slve to mgr online
The previous operation is ok ,When I execute the command "start group_replication",mysql crash.

How to repeat:
1. master-slave online change to mgr
2. 2k+ connections some insert and select  request on master
3. my.cnf :

report_host = 10.151.138.170
skip-slave-start

slave_parallel_workers = 16
slave_parallel_type = LOGICAL_CLOCK
slave_preserve_commit_order=1
master_info_repository=TABLE
relay_log_info_repository=TABLE
relay_log_recovery=ON

#gtid  mgr
gtid_mode=ON
enforce_gtid_consistency=ON
master_info_repository=TABLE
relay_log_info_repository=TABLE
binlog_checksum=NONE
log_slave_updates=ON
binlog_format=ROW
transaction_write_set_extraction=XXHASH64

loose-group_replication_group_name="94defe15-c456-11e7-b4ff-a0423f30d3b7"
loose-group_replication_ip_whitelist="10.194.7.205,10.151.138.170,10.246.137.125"
loose-group_replication_start_on_boot=off
loose-group_replication_local_address= "10.151.138.170:24901"
loose-group_replication_group_seeds= "10.151.138.170:24901,10.194.7.205:24901,10.246.137.125:24901"
loose-group_replication_bootstrap_group= off
loose-group_replication_single_primary_mode=true
loose-group_replication_enforce_update_everywhere_checks=false
[22 Nov 2017 14:17] yaodong feng 
gdb file

Attachment: mysqld.7694.tm_1510554843.uid_351.sig_11.txt (text/plain), 2.89 MiB.
[7 Dec 2017 2:07] MySQL Verification Team 
Hi,

please upload full error log

thanks
Bogdan
[8 Jan 2018 1:00] Bugs System 
No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".