Cannot Start Galera Cluster After Crash

[Afnan Ngathour Rahman]

We have a galera cluster that has 2 nodes. Yesterday the disk on node1 was full and the cluster was down, one of our team changed “wsrep_on” to “OFF” and ran node1 independently without the galera cluster. Now we try to run the galera cluster again but it always fails.

What do we do:
1. Stop mariadb in node1 and node2
2. Change safe_to_bootstrap: 1
3. Start galera cluster in node1 using “sudo galera_new_cluster” or “sudo service mysql start --wsrep-new-cluster” and get this output: “$ sudo galera_new_cluster Job for mariadb.service failed because a fatal signal was delivered causing the control process to dump core. See “systemctl status mariadb.service” and “journalctl -xeu mariadb.service” for details”

Our MariaDB version
mysql  Ver 15.1 Distrib 10.6.18-MariaDB, for debian-linux-gnu (x86_64) using  EditLine wrapper

Our MySQL log is in attachment file

Our 60-galera.conf
```
[galera]
# Mandatory settings
wsrep_on                 = ON
wsrep_cluster_name       = "CUKB Galera Cluster"
wsrep_provider           = /usr/lib/galera/libgalera_smm.so
wsrep_cluster_address    = gcomm://172.21.254.45,172.21.254.46
binlog_format            = row
default_storage_engine   = InnoDB
innodb_autoinc_lock_mode = 2
wsrep_node_address       = "172.21.254.45"
wsrep_node_name          = "node1"

# Allow server to accept connections on all interfaces.
bind-address = 0.0.0.0

# Optional settings
#wsrep_slave_threads = 1
#innodb_flush_log_at_trx_commit = 0
```

Our 50-server.conf in attachment file

Our /var/lib/mysql/grastate.dat file:
```
# GALERA saved state
version: 2.1
uuid:    45a6c7d3-52e2-11ee-8eab-1f971ac4ad6e
seqno:   -1
safe_to_bootstrap: 1
```

[Afnan Ngathour Rahman]

Resolved:

This issue related to :

https://jira.mariadb.org/browse/MDEV-27622?jql=text%20~%20%22std%3A%3Abad_alloc%22

Just add  

wsrep_provider_options="gcache.recover=off;" 
And start the cluster