Details
-
Bug
-
Status: Open (View Workflow)
-
Major
-
Resolution: Unresolved
-
10.6
-
Linux using systemd
Description
I use "galera_new_cluster mariadb@node1" to bootstap a new cluster.
If set "wsrep_cluster_address="gcomm://" in config file, it is ok to startup a new cluster,
If set "wsrep_cluster_address="gcomm://10.10.10.1:4567,10.10.10.2:4567,10.10.10.3:4567" in config file, it will fail startup a new cluster.
I find a way to fix this problem:
edit "/usr/lib/systemd/system/mariadb@.service" file and append $_WSREP_NEW_CLUSTER in the end of "ExecStart="
then mariadb@.service can catch the environment variable _WSREP_NEW_CLUSTER="--wsrep-new-cluster" (setting in "galera_new_cluster" script)
After a "systemctl daemon-reload", then "galera_new_cluster mariadb@node1" will work fine to startup a new cluster.
Attachments
Activity
| Field | Original Value | New Value |
|---|---|---|
| Description |
I use "galera_new_cluster mariadb@node1" to bootstap a new cluster.
If set "wsrep_cluster_address="gcomm://" in config file, it is ok to startup a new cluster, If set "wsrep_cluster_address="gcomm://10.10.10.1:4567,10.10.10.2:4567,10.10.10.3:4567" in config file, it will fail startup a new cluster. I find a way to fix this problem is by edit "/usr/lib/systemd/system/mariadb@.service" file and append $MYSQD_OPTS in the end of "ExecStart=". After a "systemctl daemon-reload", then "galera_new_cluster mariadb@node1" will work fine to startup a new cluster. |
I use "galera_new_cluster mariadb@node1" to bootstap a new cluster.
If set "wsrep_cluster_address="gcomm://" in config file, it is ok to startup a new cluster, If set "wsrep_cluster_address="gcomm://10.10.10.1:4567,10.10.10.2:4567,10.10.10.3:4567" in config file, it will fail startup a new cluster. I find a way to fix this problem: edit "/usr/lib/systemd/system/mariadb@.service" file and append $_WSREP_NEW_CLUSTER in the end of "ExecStart=" then mariadb@.service can catch the environment variable _WSREP_NEW_CLUSTER="--wsrep-new-cluster" (setting in "galera_new_cluster" script) After a "systemctl daemon-reload", then "galera_new_cluster mariadb@node1" will work fine to startup a new cluster. |
| Assignee | Jan Lindström [ JIRAUSER53125 ] |
| Fix Version/s | 10.6 [ 24028 ] |