[
https://issues.apache.org/jira/browse/SOLR-14636?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17154967#comment-17154967
]
Mark Robert Miller commented on SOLR-14636:
-------------------------------------------
I used up my focus hours to get here, it's a friggen rubik's cube, so a little
slower pace for a bit.
This branch is the result of a process that came out of the culmination of all
of my work on Lucene, Solr, and SolrCloud over the past 15 years.
The mistake is to think the current state of affairs can be addressed issue by
issue. It's not a stupid mistake, it's a super common one. But it's die on the
hamster wheel mistake.
The current state of affairs can be addressed though.
I like to call it "speed is the light". There is an assumption that the
SolrCloud tests in particular are slow by nature. You have ZK, god ... hdfs,
Jetty, RRDB, 5 kitechen sinks, 3 more in the guest house, and like a bajillion
3rd party libs. The assumption is wrong though. Each test has the potential to
fly.
So make the tests fly. And it's a laborious process, because the code base is
old and sprawling. But I stopped caring if I could find my way out of these
excursions long ago, so I make the tests fly one by one. And the system falls
apart. Because the system is built to survive a much more forgiving world and
it is full of gremlins and bugs and really hideous stuff to look at. It thrives
in this world where it can cause chaos and behave like a little black box of
alchemy. I squeeze with my tools right down on those tests, and I fix the
problems that start to so easily emerge. And the system starts to work. I
prefer it about 10-1000x myself.
I'll have to spend some time doing more hardening, renabling tests, and then I
have to add some new tests, that start by adding one replica, one shard, one
document. And then moving up, little by little. Finds great sh$$#.
> Provide a reference implementation for SolrCloud that is stable and fast.
> -------------------------------------------------------------------------
>
> Key: SOLR-14636
> URL: https://issues.apache.org/jira/browse/SOLR-14636
> Project: Solr
> Issue Type: Task
> Security Level: Public(Default Security Level. Issues are Public)
> Reporter: Mark Robert Miller
> Assignee: Mark Robert Miller
> Priority: Major
>
> SolrCloud powers critical infrastructure and needs the ability to run quickly
> with stability. This reference implementation will allow for this.
> *location*: [https://github.com/apache/lucene-solr/tree/reference_impl]
> *status*: alpha
> *tests***:
> * *core*: passing with ignores (not solid*)
> * *solrj*: tbd
> * *test-framework*: tbd
> * *contrib/analysis-extras*: tbd
> * *contrib/analytics*: tbd
> * *contrib/clustering*: tbd
> * *contrib/dataimporthandler*: tbd
> * *contrib/dataimporthandler-extras*: tbd
> * *contrib/extraction*: tbd
> * *contrib/jaegertracer-configurator*: tbd
> * *contrib/langid*: tbd
> * *contrib/prometheus-exporter*: tbd
> * *contrib/velocity*: tbd
> _* Running tests quickly and efficiently with strict policing will more
> frequently find bugs and requires a period of hardening._
> _** Non Nightly currently, Nightly comes last._
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
