On 22/03/19 15:02 +0300, Andrei Borzenkov wrote: > On Fri, Mar 22, 2019 at 1:08 PM Jan Pokorný <jpoko...@redhat.com> wrote: >> Also a Friday's idea: >> Perhaps we should crank up "how to ask" manual for this list > > Yest another one? > > http://www.catb.org/~esr/faqs/smart-questions.html
That's the essence that can be linked along (I had that document somewhere back on my mind when proposing that), but I also had some very specific points to raise: - decision diagram on what information is usually requested for particular community assisted debugging of particular behaviours of particular parts of the stack, possibly generalized around how to find out which of the component needs this level of investigation in the first place - link to community-managed index of notable posts previously circulated on the lists (I am not very good and possibly also biased at maintaining https://wiki.clusterlabs.org/wiki/Lists%27_Digest, anyone is welcome to step in and grow the insightful usefulness of some, sometimes very elaborate, posts to the MLs, it would certainly use some more love) - centralized summary of where the issue reports for particular components are expected, and how the existing ones can be searched - (whatever will help to cut the collateral time spent on progress towards the satisfaction of the original poster, be it a providing appropriate pieces of information off the bat, getting to understand the shared terminology to avoid communication problems like in this very case, issue routing or whatever else; still someone not getting the importance of fencing?) This is not to discourage rather a generic purposing of the lists, but primarily to limit signal-to-noise and redundancy ratio at least a bit. It's fully acknowledged that many mysterious behaviours are just a sum of the behaviours of the underlying components, and sometimes the web search will fail for some, so it's entirely OK bring up such points up, but even better if done so in a somewhat considerate way. User friendliness of the HA components is also relevant, so the better the diagnostics and documentation are to realize the mechanisms involved and their causal principles, the less question marks are to arise, sure, and that's why we need to work also on that front, it's admittedly not a one way street (and I don't want to sound harsh towards the audience, especially since I can also be blamed for contributing to the current state there -- I'd just prefer it if the lists could work at some productive level basis, since this is not exactly a hot-line service). -- Jan (Poki)
pgpAmOkqNretN.pgp
Description: PGP signature
_______________________________________________ Manage your subscription: https://lists.clusterlabs.org/mailman/listinfo/users ClusterLabs home: https://www.clusterlabs.org/