[chef] Re: Re: is there a timeout attr for the service resource?

On Fri, 13 Sep 2013, Charles Johnson wrote:

> While it's not a timeout change, what you can do instead is set a retry.
>
> All resources support a retries attribute, which takes an int as number of 
> retries to attempt before marking the resource failed. There's also a 
> retry_delay attribute, to set a timer between retries.
>
> More info here: http://docs.opscode.com/chef/resources.html#attributes

I thought about that, but thought that might not work because the "start" is
already still in flight, waiting for innodb recovery to finish. I assumed
a retry of "start" wouldn't work, might exit non-zero. But I haven't tried it
yet. I'll give it a try. Thanks.

Still, I would think a timeout on service resource would be useful.

And I guess I could also try starting mysql without the service resource, but
within a ruby_block that creates the timeout..


> On September 13, 2013 at 2:58:58 PM, 
> 
 
>  
> (
 )
>  wrote:
>
> hai.
>
> In a recipe, the service resource to start mysql threw an exception because 
> it
> took too long for it to start mysql. The node that runs this recipe should 
> be
> expected to possibly take several minutes to start mysql, since innodb is 
> doing
> a recovery.
>
> [2013-09-13T21:37:46+00:00] DEBUG: service[mysql] supports status, running
> [2013-09-13T21:47:56+00:00] DEBUG: Re-raising exception: 
> Mixlib::ShellOut::CommandTimeout - service[mysql] (dbci::drone line 61) had 
> an error: Mixlib::ShellOut::CommandTimeout: command timed out:
> ---- Begin output of /sbin/service mysql start ----
> STDOUT: Starting MySQL (Percona Server)............
>
> After this exception was thrown, mysql actually did finish successfully 
> starting
> up.
>
> I looked for a way to set a timeout for the service resource, but didn't see
> a way to do it. Can anyone advise?
>
> thanks,
> kallen