[ 
https://issues.apache.org/jira/browse/SOLR-656?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Hoss Man resolved SOLR-656.
---------------------------

       Resolution: Fixed
    Fix Version/s:     (was: 4.2)
                       (was: 5.0)
                   4.0

bq. My patch is unnecessary on 4x, all that remains for this issue is to decide 
whether the 3.6 behavior is considered a bug.

given that doing an "rm index/\*" has never been supported, and the only reason 
this issue was opened was to try and provide a better error message, i don't 
think we need to worry about fixing this in a future 3.6.x release.

bq. I am still curious whether adding isFSBased() to DirectoryFactory is 
something that holds interest. The patch is ready - all that you'd have to do 
is remove the part that wasn't necessary for this issue.

I'm not sure that i grasp the value/use of a method like that -- i would 
suggesting bring up the usecases you have in mind on the dev@lucene
                
> better error message when "data/index" is completely empty
> ----------------------------------------------------------
>
>                 Key: SOLR-656
>                 URL: https://issues.apache.org/jira/browse/SOLR-656
>             Project: Solr
>          Issue Type: Wish
>            Reporter: Hoss Man
>            Assignee: Hoss Man
>             Fix For: 4.0
>
>         Attachments: SOLR-656.patch, SOLR-656-rmdir.patch, 
> SOLR-656-rmdir.patch
>
>
> Solr's normal behavior is to create an "index" dire in the dataDir if one 
> does not already exist, but if "index" does exist it is used as is, warts and 
> all ... if the index is corrupt in some way, and Solr can't create an 
> IndexWriter or IndexReader that error is propagated up to the user.
> I don't think this should change: Solr shouldn't attempt to do anything 
> special if there is a low level problem with the index, but something that 
> i've seen happen more then a few times is that people unwittingly "rm 
> index/*" when they should "run -r index" and as a result Solr+Lucene gives 
> them an error instead of just giving them an empty index
> when checking if an existing index dir exists, it would probably be worth 
> while to add a little one line sanity test that it contains some files, and 
> log a warning.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to