You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@jackrabbit.apache.org by sowmi <so...@telenav.com> on 2006/07/25 20:58:42 UTC
Increase in save time
I am trying to load items in batches of 10000 to my repository. While the
first loads are quick enough, the save time seems to be increasing with each
subsequent batch. These are the times in seconds for each 10000 items:
42,121,186,267,309,343,372,412,451,498,524,540
Any idea why this happens? Is the save time dependent on the number of items
in repository? I am using the same session object for each of the batch
saves. Will it make a difference if I recycle the session? Please help.
--
View this message in context: http://www.nabble.com/Increase-in-save-time-tf1999932.html#a5491201
Sent from the Jackrabbit - Users forum at Nabble.com.
Re: Increase in save time
Posted by Stefan Guggisberg <st...@gmail.com>.
On 7/25/06, sowmi <so...@telenav.com> wrote:
>
> I am trying to load items in batches of 10000 to my repository. While the
> first loads are quick enough, the save time seems to be increasing with each
> subsequent batch. These are the times in seconds for each 10000 items:
>
> 42,121,186,267,309,343,372,412,451,498,524,540
>
> Any idea why this happens? Is the save time dependent on the number of items
> in repository? I am using the same session object for each of the batch
> saves. Will it make a difference if I recycle the session? Please help.
run the test with jackrabbit's default configuration (i.e.
DerbyPersistenceManager)
and compare the results. since you're using a custom persistence
manager / schema
i suspect this to be the reason for the decreasing performance you experience.
jackrabbit's scalability is mainly affected by the choice of
persistence manager/schema.
DerbyPersistenceManager should scale reasonably well.
however, if your nodes have 'large' numbers of child nodes (e.g. >
30k) performance will
be negatively affected.
cheers
stefan
>
> --
> View this message in context: http://www.nabble.com/Increase-in-save-time-tf1999932.html#a5491201
> Sent from the Jackrabbit - Users forum at Nabble.com.
>
>
Re: Increase in save time
Posted by sowmi <so...@telenav.com>.
Yes, I am using my extension of the SimpleDbPersistenceManager. I have custom
implementation of the store(NodeState) and store(PropertyState), which store
data in my own schema. They follow the same semantics of creating
PreparedStatements on initilialization, as in DbPersistenceManager.
--
View this message in context: http://www.nabble.com/Increase-in-save-time-tf1999932.html#a5493418
Sent from the Jackrabbit - Users forum at Nabble.com.
Re: Increase in save time
Posted by Vikas Bhatia <vi...@gmail.com>.
What kind of repository structure are you using. PersistenceManager?
On 7/25/06, sowmi <so...@telenav.com> wrote:
>
>
> I am trying to load items in batches of 10000 to my repository. While the
> first loads are quick enough, the save time seems to be increasing with
> each
> subsequent batch. These are the times in seconds for each 10000 items:
>
> 42,121,186,267,309,343,372,412,451,498,524,540
>
> Any idea why this happens? Is the save time dependent on the number of
> items
> in repository? I am using the same session object for each of the batch
> saves. Will it make a difference if I recycle the session? Please help.
>
> --
> View this message in context:
> http://www.nabble.com/Increase-in-save-time-tf1999932.html#a5491201
> Sent from the Jackrabbit - Users forum at Nabble.com.
>
>