[
https://issues.apache.org/jira/browse/SOLR-5652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13878020#comment-13878020
]
Hoss Man commented on SOLR-5652:
--------------------------------
bq. The next steps I can think of: ...
Done.
It also occurred to me that even though the test is just looping over the all
the sort fields, and even though the requests are *suppose* to be completely
independent of each other; the ordering of the sort fields _was_ based on a Map
iteration -- which might explain why the seeds don't reproduce for other people
on other OSs/JVMs (presuming that there is a bug in the code itself, and that
bug is dependent on some past behavior). So in addition to the extra logging i
mentioned wanting to add, the code that gets the list of all sort field names
now explicitly sorts those names, then shuffles them using the tests
{{random()}} ... so hopefully the failures will now be more easy to reproduce.
> Heisenbug in DistribCursorPagingTest: "walk already seen ..."
> -------------------------------------------------------------
>
> Key: SOLR-5652
> URL: https://issues.apache.org/jira/browse/SOLR-5652
> Project: Solr
> Issue Type: Bug
> Reporter: Hoss Man
> Assignee: Hoss Man
> Attachments: jenkins.thetaphi.de_Lucene-Solr-4.x-MacOSX_1200.log.txt,
> jenkins.thetaphi.de_Lucene-Solr-4.x-MacOSX_1217.log.txt
>
>
> Twice now, Uwe's jenkins has encountered a "walk already seen ..." assertion
> failure from DistribCursorPagingTest that I've been unable to fathom, let
> alone reproduce.
> Using this as a tracking issue to try and make sense of it.
> Summary of things noticed so far (in 2 failures):
> * So far only seen on http://jenkins.thetaphi.de
> * So far only seen on MacOSX
> * So far only seen on branch 4x
> * So far seen on both Java6 and Java7
> * fails occured in first block of randomized testing:
> ** we've indexed a small number of randomized docs
> ** we're explicitly looping over every field and sorting in both directions
> * fails were both when sorting on one of the "\*_dv_last desc" fields
> (docValues=true, sortMissingLast=true)
> ** sort on same field asc has always worked fine just before this (fields are
> in arbitrary order, but "asc" always tried before "desc")
> ** sorting on some other random fields has sometimes been tried before this
> and worked
> (specifics of each failure seen in the wild recorded in comments)
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
