On Thu, 06 May 2010 07:27:22 +1000, Ken McDonell wrote:
kenj> On Wed, 2010-05-05 at 20:09 +1000, Max Matveev wrote:
>> On Wed, 05 May 2010 07:39:30 +1000, Ken McDonell wrote:
kenj> 2. the routines are not used by pmchart
>> Which is probably a fault of pmchart, neh?
kenj> No ... if one separates the UI event-driven goo from the sync PCP glop
kenj> things tend to "just work".
As gnb has said, if there a sudden drop in TCP or if pmcd on one of
the hosts takes a bit too long to extract metrics then it just stops
working and attempts to catch up later by skipping all hosts just to
cater for a slow one.
kenj> I have a philosophical objection to bloat ... so if we're including code
kenj> that no one is using it should be removed ... in this case it seems that
kenj> there are some pockets of use outside the open source PCP code so the
kenj> precondition is not met.
Speaking of bloat - I've got a unfinished wireshark dissector written
in Lua which can grok PCP protocol and display if nicely when added to
Lua-enabled wireshark and I'm looking for a home for it. Should I park
it somewhere in the pcp tree or is there a better place for it?
kenj> But if we're going to keep the code then having no QA coverage is a very
kenj> large risk. The recent derived metrics and dynamic metrics in the PMNS
kenj> involved lobotomizing big chunks of pmns.c, and if the async stuff still
kenj> works or works at all with derived and dynamic metrics that will be good
kenj> luck rather than good management.
kenj> So if the code is going to stay, I suggest it is up to those who are
kenj> dependent on the async routines to pitch in and offer some assistance to
kenj> expand the QA coverage (at least from a selfish point of view for the
kenj> parts of the functionality they care about or depend upon).
I'll see what I can do in this area once the current leg of the death
march is over.
I was sure there were an example program which was using async API but
maybe it was stuck in some dusty non-PCP tree inside SGI, so worse
case I'd have to write it again. And the best case may be that I can
bolt in to pmchart.
Now, where was that sword again...
max
|