Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
10.0.14-galera, 10.0.15-galera
-
None
-
CentOS 6 x86_64
Description
The /etc/init.d/mysql init script returns a 0 (indicating no error) code on failure when using the bootstrap argument, used for bootstrapping a Galera cluster.
I observed this when I was running
/etc/init.d/mysql bootstrap
|
and
/usr/bin/mysqld_safe --wsrep-recover
|
at the same time
To be clear, the bug here is not that it failed in my example, but that it failed and returned 0.
[root@db01 ~]# /etc/init.d/mysql bootstrap
|
Bootstrapping the clusterStarting MySQL... [FAILED]
|
[root@db01 ~]# echo $?
|
0
|
This signals to other tools such as Puppet, Pacemaker or systemd that the MariaDB daemon is running ok while it isn't.
Attachments
Activity
| Field | Original Value | New Value |
|---|---|---|
| Assignee | Nirbhay Choubey [ nirbhay_c ] |
| Description |
I observed this command {code}/etc/init.d/mysql bootstrap{code} failing when I was running {code}/usr/bin/mysqld_safe --wsrep-recover{code} at the same time The same kan be observed when trying to bootstrap while running {code}/usr/bin/mysql start{code} at the same time. The bug here is not that it fails (it should), but that *it fails and returns 0*. {code} [root@db01 ~]# /etc/init.d/mysql bootstrap Bootstrapping the clusterStarting MySQL... [FAILED] [root@db01 ~]# echo $? 0 {code} This signals to other tools (in my case, Puppet) that the bootstrap did not fail. |
The /etc/init.d/mysql init script returns a 0 (indicating no error) code on failure when using the bootstrap argument, used for bootstrapping a Galera cluster. I observed this when I was running {code}/etc/init.d/mysql bootstrap{code} and {code}/usr/bin/mysqld_safe --wsrep-recover{code} at the same time To be clear, the bug here is not that it failed in my example, but that *it failed and returned 0*. {code} [root@db01 ~]# /etc/init.d/mysql bootstrap Bootstrapping the clusterStarting MySQL... [FAILED] [root@db01 ~]# echo $? 0 {code} This signals to other tools such as Puppet, Pacemaker or systemd that the MariaDB daemon is running ok while it isn't. |
| Summary | Galera bootstrap might fail with exit 0 if another mysql process is active. | Galera bootstrap (/etc/init.d/mysql bootstrap) returns code 0 on failure |
| Component/s | Galera [ 10124 ] | |
| Fix Version/s | 5.5.41-galera [ 17502 ] | |
| Fix Version/s | 10.0.16-galera [ 18101 ] | |
| Resolution | Fixed [ 1 ] | |
| Status | Open [ 1 ] | Closed [ 6 ] |
| Component/s | wsrep [ 11500 ] | |
| Component/s | Galera [ 10124 ] |
| Workflow | MariaDB v2 [ 58953 ] | MariaDB v3 [ 65623 ] |
| Workflow | MariaDB v3 [ 65623 ] | MariaDB v4 [ 148622 ] |
A fix for this issue would be to append
exit $?to this line.
I don't know how to use bzr and going through the contribution guidelines seems overkill for adding one statement to a shell script. Can a regular contributor add this?