Bug #109262 when communicationStack=mysql,then rebootClusterFromCompleteOutage will hang

• Submitted: 1 Dec 2022 14:34
• Modified: 6 Dec 2022 15:45
• Reporter: mosaic wang
• Status: Verified
• Category: MySQL Cluster: Replication
• Severity: S2 (Serious)
• Version: 8.0.30
• OS: CentOS (centos 7.9.2009)
• CPU Architecture: x86 (x86-64)
• Tags: rebootcluster communicationstack docker

[1 Dec 2022 14:34] mosaic wang

Description:
create mysql innodb clust with communicationstack is MYSQL
mysql-js:
var cluster = dba.createCluster("mycluster", {memberSslMode: 'REQUIRED',communicationStack: 'mysql',consistency: '1'})
cluster.addInstance("myica_admin:ica_admin123@mysql2:3306",{recoveryMethod: "clone"})
cluster.addInstance("myica_admin:ica_admin123@mysql3:3306",{recoveryMethod: "clone"})

then clust is ready.when shutdown all instance and restart instance:
dba.rebootClusterFromCompleteOutage("mycluster",{primary: "mysql1:3306"})
display "Restoring the Cluster 'mycluster' from complete outage..."
I wait over 30m,there is not more info
I find solution:
1.only start one instance
2.dba.rebootClusterFromCompleteOutage("mycluster",{primary: "mysql1:3306",force: true})
3.then start other instance
4.cluster is ready,all instance is join cluster

How to repeat:

1.setup mysql innodb cluster:
var cluster = dba.createCluster("mycluster", {memberSslMode: 'REQUIRED',communicationStack: 'mysql',consistency: '1'})
cluster.addInstance("myica_admin:ica_admin123@mysql2:3306",{recoveryMethod: "clone"})
cluster.addInstance("myica_admin:ica_admin123@mysql3:3306",{recoveryMethod: "clone"})

2.stop all instance and resart it
3.dba.rebootClusterFromCompleteOutage("mycluster",{primary: "mysql1:3306"})

[1 Dec 2022 14:37] mosaic wang

[mysqld]
# 设置为3G
innodb_buffer_pool_size=256m
innodb_buffer_pool_instances=8

# 信任子程序的创建者，禁止创建、修改子程序时对SUPER权限的要求
log_bin_trust_function_creators=1

# 客户端连接数
max_connections=1000

# 大小写不敏感(0:大小写敏感)
lower_case_table_names=1

## 同一局域网内注意要唯一
server_id=100

gtid_mode=ON
enforce_gtid_consistency=ON

ssl_ca=/certs/ca-cert.crt
ssl_cert=/certs/server.crt
ssl_key=/certs/server.key

# 日志采用系统时间(缺省是UTC)
log_timestamps=SYSTEM

# 提升性能
binlog_transaction_dependency_tracking=WRITESET

# 启用组复制而必须设置的参数
log_bin=mylog-bin
plugin_load_add='group_replication.so'

# 格式为:ip:端口.设置为本机IP地址
group_replication_local_address=[fd15:4ba5:5a2b:1008:192:168:132:42]:3316
group_replication_group_seeds=[fd15:4ba5:5a2b:1008:192:168:132:42]:23316,[fd15:4ba5:5a2b:1008:192:168:132:43]:23316,[fd15:4ba5:5a2b:1008:192:168:132:44]:23316

# 中继日志
relay_log=mysql-relay

[6 Dec 2022 15:45] MySQL Verification Team

Thank you for the report