Re: pcp updates: libpcp connection unsharing

----- Original Message -----
> > There may be negative testability implications, in that pmcd would
> > have to retain code for incoming multiplexed connections, but the
> > mainstream code would no longer create them, so only old (or fake)
> > clients would stress that part of pmcd.  (We should probably have
> > somewhere a suite of such old binaries to try for intercompatibility
> > with newer servers anyway.)
> 
> Ok, thanks Frank - so we need to retain connection sharing server-side for
> downrev clients. But new clients would no longer use it. We need downrev
> clients exercising QA for uprev pmcd. And similarly, QA for uprev clients
> talking to downrev pmcd. None of this QA exists .. so it would seem this
> patch isn't ready for merging yet ..?

Agreed.

> Any further comments or thoughts anyone?
> 

Hmm - AFAICT there seems to be no evidence this change is actually fixing
anything.  I'm all for code simplification, but the rationale here is not
compelling (don't get me wrong - I wrote the comment in context.c ...)

         * It is the topic of some debate as to whether PMCD connection
         * sharing is of much value at all, especially considering the
         * number of subtle and nasty bugs it has caused over time.  Do
         * not rely on this behaviour, it may well be removed someday.

We need to properly understand the root cause of the problem; e.g. I wonder
if its more pmwebds use of an async pmDestroyContext on a context that was
recently timed out by libpcp (and hence experienced a "partial" context
tear down in libpcp) - perhaps there's some completely unrelated botch in
the libpcp pmDestroyContext logic.  It would not be the first time & that's
a fairly rarely traveled code path.

If it helps any, there is an undocumented hostspec flag ("exclusive") which
can be used here - Henry and I used that when we were diagnosing an earlier
round of pmwebd timeouts/failures (using it required modifications to Vector
source code, IIRC, Henry may still have a patch handy).  For that problem we
guessed context sharing might be a cause, too, but found no improvement from
disabling it FWIW.

cheers.

--
Nathan

-=-=-=-=-=-=-=-=-=-=-=-
pcp mailing list
[email protected]
https://groups.io/g/pcp/messages
-=-=-
Groups.io Links:

You receive all messages sent to this group.

View/Reply Online (#14664): https://groups.io/g/pcp/message/14664
View All Messages In Topic (8): https://groups.io/g/pcp/topic/3038481
Mute This Topic: https://groups.io/mt/3038481?uid=174580
New Topic: https://groups.io/g/pcp/post

Change Your Subscription: https://groups.io/g/pcp/editsub?uid=174580

Group Home: https://groups.io/g/pcp
Contact Group Owner: [email protected]

Terms of Service: https://groups.io/static/tos

Unsubscribe: https://groups.io/g/pcp/leave/354243/563757577/xyzzy
-=-=-=-=-=-=-=-=-=-=-=-