2

I've got a service set up to start at boot but it's not doing so and I don't know where to look (e.g. logs) to troubleshoot it.

I've verified the run level:

$ grep default /etc/inittab 
id:3:initdefault:

And the symlinks are there:

$ find /etc -name *beans*
/etc/rc.d/rc6.d/K47beanstalkd
/etc/rc.d/rc1.d/K47beanstalkd
/etc/rc.d/rc0.d/K47beanstalkd
/etc/rc.d/rc2.d/K47beanstalkd
/etc/rc.d/rc4.d/K47beanstalkd
/etc/rc.d/rc3.d/K47beanstalkd
/etc/rc.d/rc5.d/K47beanstalkd
/etc/rc.d/init.d/beanstalkd

And when I start it manually (via "sudo /etc/init.d/beanstalkd start") it works fine.

Where do I start troubleshooting this? Where might the boot sequence be logged that I can grep for issues? This is a CentOS box if that matters.

Thanks.

Teflon Ted
  • 500
  • 2
  • 7
  • 18

1 Answers1

6

Your symlinks are there, but they're not what you want. If you notice, yours are all prefixed with 'K' for kill. They should be 'S' for start.

Fix this with chkconfig beanstalkd on and verify it with chkconfig --list | grep beanstalkd.

pboin
  • 1,086
  • 8
  • 11
  • 1
    Yep, but the S *and* K files should be there, as it needs to start on entry to the runlevel, and stop on exit. On another note: it DOES matter at little that it's CentOS/Fedora/RHEL, as Debian-like distros use update-rc.d instead of chkconfig. – Lee B Oct 22 '09 at 14:05
  • Pedants unite! You are correct -- services that are to run should have both 'S' and 'K' scripts, not just 'S'. Good catch. – pboin Oct 22 '09 at 15:18