You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@jackrabbit.apache.org by Ian Boston <ie...@tfd.co.uk> on 2007/06/15 19:03:49 UTC

Journal Maintanence Question

Hi,
Quick question,
In a FileJournal for a cluster the files get rotated and can be managed, 
however in a DB journal is their the same rotation or cleanup going 
on... I think thats right ?

In the DatabaseJournal however...

I am concerned about having millions of journal records hitting db 
performance and I was wondering if it would make more sense to have a 
policy, either

records older than the oldest cluster node will be removed or

records older than 30 minutes will be removed.

Obviously, this will require a snapshot of the local space to seed a new 
cluster node, and that will have to be fresher than the oldest journal 
entry.

Am I worrying unnecessarily ?

Ian

Re: Journal Maintanence Question

Posted by Dominique Pfister <do...@day.com>.
Hi Ian,

at the moment, records in the database journal are actually never
removed, which enables to create a completely empty repository cluster
node that will automatically synchronize to the cluster's state at any
point in the future. But, of course, introducing policies that
restrict the growth and size of the journal table, is an absolutely
reasonable enhancement, so please create a JIRA issue for this.

Kind regards
Dominique

On 6/15/07, Ian Boston <ie...@tfd.co.uk> wrote:
> Hi,
> Quick question,
> In a FileJournal for a cluster the files get rotated and can be managed,
> however in a DB journal is their the same rotation or cleanup going
> on... I think thats right ?
>
> In the DatabaseJournal however...
>
> I am concerned about having millions of journal records hitting db
> performance and I was wondering if it would make more sense to have a
> policy, either
>
> records older than the oldest cluster node will be removed or
>
> records older than 30 minutes will be removed.
>
> Obviously, this will require a snapshot of the local space to seed a new
> cluster node, and that will have to be fresher than the oldest journal
> entry.
>
> Am I worrying unnecessarily ?
>
> Ian
>