[ClusterLabs] [questionnaire] Do you manage your pacemaker configuration by hand and (if so) what reusability features do you use?

[Jan Pokorný]

On 31/05/18 14:48 +0200, Jan Pokorný wrote:
> I am soliciting feedback on these CIB features related questions,
> please reply (preferably on-list so we have the shared collective
> knowledge) if at least one of the questions is answered positively
> in your case (just tick the respective "[ ]" boxes as "[x]").

I am not sure how to interpret no feedback so far -- does it mean
that those features are indeed used only very sparsely, or is the
questionnaire not as welcoming as it could be?  This is definitely
not the last time the userbase's feedback is of help, so the more
pleasant we can do such enquiries, the better turnaround, I guess.

> Any other commentary also welcome -- thank you in advance.
> 
> 1.  [ ] Do you edit CIB by hand (as opposed to relying on crm/pcs or
>         their UI counterparts)?

Putting seriousness aside for a bit, there's a relevant anecdotical
reference directly from pacemaker's own codebase:
https://github.com/ClusterLabs/pacemaker/blob/Pacemaker-2.0.0-rc5/Makefile.common#L54
:-)

As Ken noted, crm shell may support all of 2. + 3. + 4., it was
just my extension that those could be especially handy with direct
XML-level control, as shifting towards more abstract thinking
about the configuration may actually conflict with the goal of
straightforward conceptual comprehension, at least in case of 3.

> 2.  [ ] Do you use "template" based syntactic simplification[1] in CIB?
> 
> 3.  [ ] Do you use "id-ref" based syntactic simplification[2] in CIB?
> 
> 3.1 [ ] When positive about 3., would you mind much if "id-refs" got
>         unfold/exploded during the "cibadmin --upgrade --force"
>         equivalent as a reliability/safety precaution?

This was a premature worst-case conclusion on my end (generally,
I think it's better to start pessimistically only to be pleased
later on, rather than vice-versa).  In fact, there's nothing that
would prevent reversibility of temporary limited-scope "id-refs"
exploding, in an unfold-upgrade-refold manner, sorry for the noise
(https://github.com/ClusterLabs/pacemaker/pull/1500).

However, you can take this also as discussion-worth probe into how
mere _syntactic_ changes not affecting the behaviour (i.e. the
semantics encoded with either syntactic expressions) whatsoever
would be perceived.  In this now merely theoretic case, parts of
information that only have bearing on user's comprehension would
be lost (multiple duplicate entities as opposed to shared single
point of control) and the question hence is:

How much frustration could arise from such semantics-preserving
interventions inflicted with schema upgrades or elsewhere?
Is this something we should avoid at all costs so as not to
alienate not even a single user, or is there some extent of
tolerance as long as you can hardly tell a difference in
higher-level tools?

> 4.  [ ] Do you use "tag" based syntactic grouping[3] in CIB?

The original questions are still valid, feel free to respond
to them or to the new bunch at your convenience.  It will help
to shape future directions for pacemaker.

-- 
Jan (Poki)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <https://lists.clusterlabs.org/pipermail/users/attachments/20180605/1318cf00/attachment.sig>