Return code incorrect when garbd is already running

[jmrenouard]

Context:
SysV init.d script: /etc/init.d/garb
Garbd is already running and started correctly.

Observed result:
service garb start
echo $?
0

Awaited result:
When garbd is already started, return code shouldn't return 0.
0 means success
If garbd is already running so trying to start garbd is expected to fail and return an non-zero return code.

Solution: return 3 instead of 0

--- garbd_origin 2014-04-15 16:48:20.115360380 +0200
+++ garb 2014-04-15 16:49:24.866360234 +0200
@@ -97,7 +97,7 @@

        if [ -r $PIDFILE ]; then
                log_failure "$prog is already running with PID $(cat ${PIDFILE})"
- return 0
+ return 3
        fi

        [ -x $prog ] || return 5

[jmrenouard (jmrenouard)]

I have found a similar bug around stop function in this file: garb

function stop return 0 if garbd is not started. stop function did not successed stopping an non running program so non zero return code is better.

Best regards,

[Przemek (pmalkowski)]

First of all, it does NOT seem that this is a correct practice in Linux to exit with '3' when using start on running service. See these examples:

[root@percona3 ~]# cat /etc/issue
CentOS release 6.5 (Final)

[root@percona3 ~]# service postfix status
master (pid 1278) is running...
[root@percona3 ~]# service postfix start
[root@percona3 ~]# echo $?
0
[root@percona3 ~]# service postfix status
master (pid 1278) is running...

[root@percona3 ~]# service crond status
crond (pid 1330) is running...
[root@percona3 ~]# service crond start
[root@percona3 ~]# echo $?
0

root@vagrant-debian-wheezy:~# cat /etc/issue
Debian GNU/Linux 7 \n \l

root@vagrant-debian-wheezy:~# service cron status
cron is running.
root@vagrant-debian-wheezy:~# service cron start
Starting periodic command scheduler: cron.
root@vagrant-debian-wheezy:~# echo $?
0

Also according to these documents, exit code 0 is a proper one in case on a start attempt when the service is already started:
http://refspecs.linuxbase.org/LSB_3.1.1/LSB-Core-generic/LSB-Core-generic/iniscrptact.html
http://fedoraproject.org/wiki/Packaging:SysVInitScript#Exit_Codes_for_non-Status_Actions

Exit code 3 is proper only for a "service xxx status" command when service is not running.

[Przemek (pmalkowski)]

Now it seems that the garb init script does wrong thing in both 3.x and 2.x versions:

[root@percona3 ~]# rpm -qa|grep -i garb
Percona-XtraDB-Cluster-garbd-2-2.11-1.2675.rhel6.x86_64
[root@percona3 ~]# ps -ef|grep garb
root 8868 8413 0 14:56 pts/0 00:00:00 grep garb
[root@percona3 ~]# service garb status
garbd is stopped
[root@percona3 ~]# echo $?
3
[root@percona3 ~]# service garb start
Starting /usr/bin/garbd: [FAILED]
[root@percona3 ~]# echo $?
0
[root@percona3 ~]# service garb status
garbd (pid 8905) is running...
[root@percona3 ~]# service garb start
/usr/bin/garbd is already running with PID [FAILED]
[root@percona3 ~]# echo $?
3

---------
[root@percona3 ~]# rpm -qa|grep -i garb
Percona-XtraDB-Cluster-garbd-3-3.7-1.3254.rhel6.x86_64
[root@percona3 ~]# service garb status
garbd is stopped
[root@percona3 ~]# echo $?
3
[root@percona3 ~]# service garb start
Starting /usr/bin/garbd: [ OK ]

[root@percona3 ~]# service garb status
garbd (pid 8641) is running...
[root@percona3 ~]# echo $?
0
[root@percona3 ~]# service garb start
/usr/bin/garbd is already running with PID 8641 [FAILED]
[root@percona3 ~]# echo $?
3

See also: lp:1367956

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PXC-1145