[Linux-ha-dev] Bug in lrmd / OCF RAs

[Alan Robertson]

Lars Marowsky-Bree wrote:
> On 2004-09-08T10:32:15,
>    Alan Robertson <alanr at unix.sh> said:
> 
> 
>>IIRC, when you eventually get passed timeouts, they'll also come in through 
>>the name/value pairs for the RA's benefit.  If/when this happens, this code 
>>will break even with correct CIB data.  So, it's also fragile with respect 
>>to minor (and expected) changes in the requirements.
> 
> 
> The timeouts are actually already passed in in the lrm_msg.h; the lrmd
> actually already uses it. I'm fine with this approach.
> 
> But I'm not so fond of the LRM sniffing the instance parameters. If we
> want to drive the LRM via name/value options too, we should put these
> into a separate hashtable to not have namespace clashes.
> 
> I'd say the instance parameters should be opaque to the LRM; it just
> maps them in whatever way the Resource Class plugin wants to. 
> 
> (If the LRM really wants to export it's internal attributes to the RA,
> it can again map them with a specific environment prefix and provide
> them to the RA.)

That makes sense.


-- 
     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