[OCF] ocf-shellfuncs standardization - Comments? Objections?

[Alan Robertson]

Nick Stoughton wrote:
> On Mon, 2005-07-04 at 17:49, Alan Robertson wrote:
>>Ragnar Kjørstad wrote:
>>>On Sun, Jul 03, 2005 at 04:32:36PM -0600, Alan Robertson wrote:
>>>>>http://refspecs.freestandards.org/LSB_2.1.0/LSB-Core-generic/LSB-Core-generic/iniscrptfunc.html
>>>>>
>>>>>log_success_msg, log_failure_msg and log_warning_msg.
>>>>OK.
>>>>
> See also the latest version, 3.0, at
> http://www.linuxbase.org/spec/booksets/LSB-Core-generic/LSB-Core-generic/iniscrptfunc.html
>>>>Those don't serve the same purpose as we're trying to do.
>>>>
>>>>These are only for the final "I'm done and everything is success or 
>>>>failure or warning" - and they only go to the pre-boot console (in 
>>>>practice).
>>>I can't find anything in the documentation that indicates that they are
>>>only for the final message. 
> These are functions designed exclusively for use in init scripts. If you
> are wanting a more general purpose logging function, LSB 3.0 includes
> the POSIX "logger" command, which writes to syslog.

All we're doing is init scripts (see the OCF spec on resource agents) No 
problem.  BUT, we aren't init.  And, we don't only run at 
startup/shutdown - so they way they're implemented and the "cultural 
expectations" about how they work kind of stink...

But that doesn't mean that we can't just go with what the spec says and 
ignore the historical implementations and cultural expectations.  In 
fact, I think I've decided that Ragnar is right, and that's what we 
ought to do...

>>>>We need things which also allow informative messages, debug messages, 
>>>>and which don't (in practice) print funny reverse video thing, and which 
>>>>don't go to the pre-boot console.
>>>The LSB only defines what the functions are, not what they should do.
>>>
> Well...that's overstating it more than a bit! The LSB definitely defines
> what things should do, and the 3.0 versions of the init script logging
> functions state that these functions should print a suitable message
> (doesn't say where though!)

Then that's not very useful is it?  I stand by my description - (even 
though it had a bit of hyperbole in it).


>>>If a script is executed as a regular init script, the scripts should
>>>probably forward the data to the screen, as they do with Red Hat and
>>>SUSE initscripts.
>>>
>>>If the script is executed in the OCF environment the same functions
>>>should (IMHO) be defined to send the messages to the Resource Manager,
>>>and then it would be up to the RM to decide what to do with them.
>>>
>>>I agree though, that the 3 functions should be complemented by
>>>log_info_msg and log_debug_msg (and possible others), which could be
>>>NOOPs in regular initscripts but still do something in the OCF
>>>environment. (Actually, it would be very useful if one could enable
>>>debugging for regular initscripts as well, e.g. by setting an
>>>environmentvariable LSB_DEBUG)
>>You can't find my descriptions in the LSB - because the aren't so well 
>>defined as that.
>>
> As I said, they simply cause the message to be "printed" (and in the
> implementations I've checked, they simply use "echo" to do that).
> 
>>But, in practice that's what they do, and that's what they're *expected* 
>>to do.
>>
> Yes ... and if you want more, then use the other interfaces defined in
> the spec, such as logger.

We're not going to drag in the rest of the world at this point - what 
we're trying to do is more restricted than that.

For a variety of reasons syslog doesn't work terribly well for HA 
systems (or at least not in our experience) - and every known HA system 
implements its own logging system -- none rely strictly on syslog.  I 
don't want to go down that rabbit trail that right now - so I'll try and 
stay a little focused for a minute :-)

All I'm concerned about in the context of this particular discussion is 
init-script compatibility...


>>One of the nice things about the ocf_log functions as we have 
>>implemented them is that they work a lot more like syslog - they 
>>automatically include date/time/scriptname in the output.
> If what you want is syslog, then use "logger".
>>Do you know of a good way to implement what you've suggested?  That is, 
>>to have them do whatever the local LSB implementation does when not 
>>invoked with OCF vars not set, and to have them do what the local OCF 
>>implementation wants, when they're invoked in the OCF environment.  I 
>>guess I know of one way.  Maybe you can suggest a better one?
>>
>>The way I know of is to have a dot-included file that redefines what 
>>those functions do, and then "dot-include" that file when in the OCF 
>>environment.  That'd work.  Not too elegant, but should be effective.
>>
>>And, adding an LSB_DEBUG and/or OCF_DEBUG flags.  That sounds like a 
>>good idea.
>>
>>LMB:  What do you think about this?


-- 
     Alan Robertson <alanr at unix.sh>

"Openness is the foundation and preservative of friendship...  Let me 
claim from you at all times your undisguised opinions." - William 
Wilberforce