jackrabbit-oak-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jukka Zitting <jukka.zitt...@gmail.com>
Subject Re: Reduce number of calls from Oak to Mongo DB on restarts
Date Fri, 25 Oct 2013 16:10:33 GMT

On Fri, Oct 25, 2013 at 11:32 AM, Chetan Mehrotra
<chetan.mehrotra@gmail.com> wrote:
> Thoughts?

All the suggested approaches seem reasonable.

More generally though, I think what you're seeing here is ultimately
an issue of granularity of access. You have a use case where the
client is accessing thousands of nodes. Because each node is accessed
using a separate database call, there's no way around those calls
until the nodes become cached. Persisting the cache across restarts
and otherwise improving it to avoid unnecessary cache misses of
previously loaded items will help, but won't take away the core issue
that accessing a thousand previously unseen nodes will require at
least a thousand database calls. And since most nodes are really small
(see http://markmail.org/message/kxe3iy2hnodxsghe), the call overhead
is significant.

The SegmentMK addresses this issue by grouping related content into
larger segments and thus reducing the number of network or disk
accesses needed in most use cases. Perhaps the MongoMK could do
something similar?


Jukka Zitting

View raw message