On Apr 24, 2006, at 10:38 AM, Lars Marowsky-Bree wrote:
On 2006-04-24T10:28:52, Andrew Beekhof <[EMAIL PROTECTED]> wrote:
yes, differing monitoring depths.
Hrm, I can see this. Yet, the last one should be entirely sufficient?
No. At best its a waste of resources, clutters the logs and makes it
look like we're doing way more than we really are.
At worst the "deep" monitor will have a "failed" result wiped out by
the "shallow" one that happens more frequently and didn't detect a
problem.
Its a bug, its fixed, can we move on? What is the concern here?
Nothing fundamentally _wrong_ will happen.
Are you suggesting that we should remove the patch because we wont
fall over in a screaming heap?
probes vs. a regular monitor are another case which i take advantage
of in the PE.
Again, a probe is not different in result from a regular monitor
op, no?
I don't see what you're trying to get at here.
The RA doesn't even know.
The RA doesn't need to know and this change does not affect the RA...
so why are we talking about RAs?
The LRM is required to report an accurate history of the resource
because the CRM needs to know. Thats what this is about.
It didn't and now it does.
(Oh, yes, so it knows because you pass in that detail via the repeat
interval, yet, the RA shouldn't do anything about it, and it can't
rely
on it, as that is not specified in the OCF docs ;-)
huh?
There is nothing that says a monitor op cannot have an interval of zero.
--
Andrew Beekhof
"Too much knowledge leads to confusion; Too many guitar lessons lead
to jazz-fusion!" - TISM
_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
