[ 
https://issues.apache.org/jira/browse/OAK-3066?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15121143#comment-15121143
 ] 

Vikas Saurabh commented on OAK-3066:
------------------------------------

bq. Instead of exposing the two types of cache maps I would rather try to hide 
it. AFAICS the methods nodeDocsAsMap() and previousDocsAsMap() are only used in 
the invalidate calls and only the keys are needed. Why not create a single 
method that gives you all the keys?
Actually, mongo store works on keys, while rdb works on values. I think we 
should rather expose a single full-invalidate method (maybe, the same way as 
mongo is doing) and let go of map views.

bq. Similarly I would not create explicit methods for the two types of stats, 
but rather change the return type of the existing method to return an 
Iterable<CacheStats>.
>From stats pov, it seems useful to distinguish between mem and pers cache (the 
>jmx are exposed separately anyway, so may be merging these 2 might not be a 
>bad idea).

bq. I would not reduce the share of other explicitly controlled caches, but 
rather let the builder reduce the remaining share of the document cache 
automatically. 
Ack.

> Persistent cache for previous documents
> ---------------------------------------
>
>                 Key: OAK-3066
>                 URL: https://issues.apache.org/jira/browse/OAK-3066
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: core, mongomk
>            Reporter: Marcel Reutegger
>            Assignee: Vikas Saurabh
>            Priority: Blocker
>              Labels: performance
>             Fix For: 1.4, 1.3.15
>
>         Attachments: OAK-3066.take1.patch, OAK-3066.wip.patch
>
>
> Previous (aka split) documents contain old revisions and are immutable 
> documents. Those documents should go into the persistent cache to reduce 
> calls to the underlying DocumentStore.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to