[infinispan-issues] [JBoss JIRA] (ISPN-4584) Stricter validation of cache configurations for distributed indexes

[Sanne Grinovero (JIRA)]

    [ https://issues.jboss.org/browse/ISPN-4584?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12988889#comment-12988889 ] 

Sanne Grinovero commented on ISPN-4584:
---------------------------------------

More short-hand and automated configuration should be delivered in ISPN-4340, here I'm literally exploring what we can do to help debugging for the power users who actually want to do the fine tuning.

The names are inferred already (by default) but we allow to override them, and the test in ISPN-4577 was overriding them.
Same for replTimeout and lock acquisition timeout, those values were set as they were for for reasons unrelated to the indexing system.

Also there is a valid short hand for the property:
{code}<property name="default.indexmanager" value="infinispan" />{code}
Technically this is the only required option, as even lucene_version was not set to the ideal version (but acceptable).

Indeed the metadata Cache could be forced to require REPL, but that's unless it's single node in which you want LOCAL, or unless it's managing independent indexes per node.. tricky. The data storing cache also should be allowed to be anything :-/

So the problem is literally that an advanced user can override most of this and shoot himself. Another problem is that the exceptions I was getting from the failing test is very misleading as there is a long chain of "caused by.." and the interesting part is swallowed: the user facing error is the infamous "UserType is not annotated with @Indexed" just that in this case it was a flat classloader, embedded mode and definitely annotated.

(What was happening is that the EOF triggered by the missing bytes in the storage was happening during the second mutation of the SearchFactory, in which case the type is already marked as "installed" into the known types registry, but the SearchFactory had actually rejected the mutation).


> Stricter validation of cache configurations for distributed indexes
> -------------------------------------------------------------------
>
>                 Key: ISPN-4584
>                 URL: https://issues.jboss.org/browse/ISPN-4584
>             Project: Infinispan
>          Issue Type: Enhancement
>      Security Level: Public(Everyone can see) 
>          Components: Lucene Directory
>            Reporter: Sanne Grinovero
>            Assignee: Gustavo Fernandes
>            Priority: Minor
>
> See also ISPN-4577 : it should not be allowed to configure a distributed metadata cache while the chunks cache is using local mode.
> Ideally think of additional strict checks which we should apply.. suggestions?
> Mitigated by ISPN-4340



--
This message was sent by Atlassian JIRA
(v6.2.6#6264)