Re: First class support for node roles

2021-11-23 Thread Ishan Chattopadhyaya
I've synced back all changes to confluence. Both represent the same document as of now. Please review and suggest if there are any outstanding concerns. Thanks for all the feedback. I wish to proceed with the implementation based on lazy consensus again. On Tue, Nov 23, 2021 at 9:38 PM Mike Drob

Re: Reference guide for 8.11

2021-11-23 Thread Cassandra Targett
Just to close the loop, I’m a little later than my promised last Friday for fixing updating the missing 8.11 notes and the problem with the 8.10 Upgrade notes page still showing DRAFT watermark, but I fixed both yesterday. Thanks for the kind words in general - always very nice to hear. I fully

Re: First class support for node roles

2021-11-23 Thread Mike Drob
I should clarify, that my -1 was specifically about reaching lazy consensus and not about that proposal itself. On Tue, Nov 23, 2021 at 9:42 AM Ishan Chattopadhyaya < ichattopadhy...@gmail.com> wrote: > > -1, I would like to see a proposal on list where it will be permanently > available > > It w

Re: First class support for node roles

2021-11-23 Thread Ishan Chattopadhyaya
> -1, I would like to see a proposal on list where it will be permanently available It will be available in the confluence document (as it currently is). The google docs is just for collecting temporary feedback, thanks to easy inline commenting capability. We promised to consolidate feedback in t

Re: First class support for node roles

2021-11-23 Thread Ishan Chattopadhyaya
> -1, I would like to see a proposal on list where it will be permanently available > in the archives instead of being directed to a Google doc which can be edited at any point in time. > This has been an incredibly difficult proposal to keep track of. This entire SIP process has been incredibly d

Re: First class support for node roles

2021-11-23 Thread Ishan Chattopadhyaya
Okay, sure, I'll rescind my assertion about having reached a lazy consensus. I'll work through all the recent comments and feedback. I think at the last moment, there was an edit by Noble to the Google Doc that hasn't made it into the SIP document, and I'll review that change, the recent comments a

Re: First class support for node roles

2021-11-23 Thread Gus Heck
Also, I don't know that it's defined anywhere, but I feel that declaring lazy consensus on a SIP in less than a week is rushing things. Some folks won't have the flexibility to address things daily and I've been uncomfortably pulled away from paid work trying to keep up with this as is. Lazy consen

Re: First class support for node roles

2021-11-23 Thread Gus Heck
IMHO two things must happen before lazy consensus wait period can begin: 1) The wiki must be updated to reflect the results of discussion in the google doc. 2) The fact that the wiki has been updated, and that you think the discussion is at an end must be posted here, preferably with a summary

Re: First class support for node roles

2021-11-23 Thread Mike Drob
-1, I would like to see a proposal on list where it will be permanently available in the archives instead of being directed to a Google doc which can be edited at any point in time. This has been an incredibly difficult proposal to keep track of. Mike On Tue, Nov 23, 2021 at 6:09 AM Ishan Chatto

Re: First class support for node roles

2021-11-23 Thread Jan Høydahl
There is still discussion and disagreement in the Google doc. The Google doc proposes that zk role structure is 1:1 for all roles -- except the "data" role. I assume the reason for that choice, even if not spelled out in the doc, is to optimize for large clusters with 1000 nodes, and thus avoid

Re: First class support for node roles

2021-11-23 Thread Ishan Chattopadhyaya
This proposal has passed with lazy consensus. We can proceed to the implementation phase. Thanks to everyone for feedback, esp. Gus for the patience. On Sun, Nov 21, 2021 at 2:24 PM Gus Heck wrote: > Thanks for adding that :) and Thanks to Nobel for translating my ravings > :). Sorry about the a