You are viewing a plain text version of this content. The canonical link for it is here.

Posted to oak-dev@jackrabbit.apache.org by Julian Reschke <ju...@gmx.de> on 2014/10/25 13:08:25 UTC

DocumentNodeStore backgroundRead

Hi again.

I've been looking at what cache misses the RDBDocumentStore gets.

It seems that the DocumentNodeStore's background threads run once per 
second by default.

The first thing the backgroundRead does is:

>         String id = Utils.getIdFromPath("/");
>         NodeDocument doc = store.find(Collection.NODES, id, asyncDelay);

where asyncDelay is 1000 as well.

This causes the find on 0:/ to produce a cache miss, unless there was 
other activity on that cache entry within the last 1000ms, thus causing 
a roundtrip to the persistence.

Is this really the intention?

Best regards, Julian

Re: DocumentNodeStore backgroundRead

Posted by Thomas Mueller <mu...@adobe.com>.

Hi,

Changes from other cluster nodes are detected that way (if a cluster node
changes anything, it writes to the root node once a second).

Regards,
Thomas

On 25/10/14 13:08, "Julian Reschke" <ju...@gmx.de> wrote:

>Hi again.
>
>I've been looking at what cache misses the RDBDocumentStore gets.
>
>It seems that the DocumentNodeStore's background threads run once per
>second by default.
>
>The first thing the backgroundRead does is:
>
>>         String id = Utils.getIdFromPath("/");
>>         NodeDocument doc = store.find(Collection.NODES, id, asyncDelay);
>
>where asyncDelay is 1000 as well.
>
>This causes the find on 0:/ to produce a cache miss, unless there was
>other activity on that cache entry within the last 1000ms, thus causing
>a roundtrip to the persistence.
>
>Is this really the intention?
>
>Best regards, Julian