Hi, I have to disagree. In contrast to Elasticsearch, most installations I have seen for small and medium sized companies were standard single or master/slave Solr installations. The reason for that was not complexity, just that if you have a standard online shop, the number of products is only millions (and that's large already). No need for shards and a single master for indexing is perfectly fine. For searching those use the master, too or they create plain old replicas. The main reason for this setup is quite clear: have all indexing on one node and use others for searching.
So please keep this simple setup! Uwe Am May 23, 2021 5:59:45 PM UTC schrieb Alexandre Rafalovitch <[email protected]>: >The worry I have is that this (and previous cloud-first) move pushes >Solr - even further - into Innovator's Dilemma situation. It is a >complicated and nuanced worry, so I would understand if nobody else >shares it. But I felt it was still worth mentioning. > >The core issue is that even though Solr is a great fit for smaller >projects with a standalone mode, we are increasingly making it more >and more complicated to understand/discover. The tutorials switching >to cloud mode was one of those "straws". This proposes to be another. > >Which means the people will just assume Solr is for gigantic projects >and avoid it all together, shrinking our total ecosystem even further. >I see that already in public discussions, whenever Solr is mentioned. > >Regards, > Alex. >P.s. I also feel that Schrodinger's "Reference branch" is also a >factor in this discussion, but that is an even messier point to try >putting in words > in a balanced way. >P.p.s. Also, when we say "docker container could run in cloud mode", >do we still mean a single-node with built-in Zookeeper (sharing Jan's >worry with that) or something that actually rolls-out 2x+1 Zookeeper >nodes, 2+ Solr nodes, etc as a "first glimpse" setup? > >On Fri, 21 May 2021 at 20:00, Jan Høydahl <[email protected]> >wrote: >> >> I'd be all for it. >> >> Possible side effect is that people perhaps get an impression that >the built-in Zookeeper is production ready. >> Another side effect is that users that on purpose run Solr standalone >will now need to run with -l flag to avoid starting in coud mode by >default. But that's their fault if they don't read the upgrade notes of >course... >> >> Jan >> >> 21. mai 2021 kl. 17:28 skrev Houston Putman <[email protected]>: >> >> Hey everyone, >> >> Currently when running bin/solr, and therefore when using the Solr >Docker image, if you start solr without specifying any options, it will >start in Standalone mode. However, many of the tutorials and examples >we give are for Solr Cloud. And there are useful features for new users >that are not supported in standalone mode, including the new >schema-designer UI. I think it would be good for new users to the >project to use the generally recommended running mode (cloud) by >default. >> >> Clearly this won't be done in 8.x. I think it would be ok to make the >switch in Solr 9, but maybe it's something we aim for Solr 10 instead. >In the meantime we should probably include a backwards-compatible >standalone flag (-l/-standalone) that will start to be used once the >default mode is switched. >> >> There has been some discussion about deprecating standalone Solr, I >want to emphasize this is not at all part of the discussion of this >thread. >> >> - Houston >> >> > >--------------------------------------------------------------------- >To unsubscribe, e-mail: [email protected] >For additional commands, e-mail: [email protected] -- Uwe Schindler Achterdiek 19, 28357 Bremen https://www.thetaphi.de
