Hi,

the point is when the database is clustered, it's split up in many
databases really with long IDs in their names, so it's not just "use the
backdoor port". Do an _all_dbs on the backdoor port to find out the exact
names.
I never tried to compact them manually - you may want to backup before
trying. Btw: there is a discussion already going on:
http://mail-archives.apache.org/mod_mbox/couchdb-dev/201505.mbox/%3C20150513202728.BF32125093%40wilderness.apache.org%3E

Good luck - sorry there is no silver bullet here. Would be great when you
can let us know your result / experience!
     Sebastian

On Fri, May 15, 2015 at 10:14 PM, Hristo Trendev <[email protected]>
wrote:

> Hi,
>
> I've tried this one as well, but the error was:
>
> {"error":"not_found","reason":"no_db_file"}
>
> It seems that the 5986 endpoint doesn't know anything about the database.
> If compaction needs to be done per-shard, as the error message says, then
> I'd expect to see something that identifies the shard to be compacted. Like
> 00000000-1fffffff for example.
>
> On Fri, May 15, 2015 at 9:00 PM, Alexander Shorin <[email protected]>
> wrote:
>
> > > I have also read in an email on the dev list, that this needs to be run
> > against the 5986 port, but again no clue what the exact URL is. I have
> > tried to guess it, but none of the combinations that I've tried did work.
> >
> > curl -XPOST http://localhost:5986/<db>/_compact
> >
> > Cluster-wide compaction is not supported yet, so you have to run it
> > per node basis.
> >
> > --
> > ,,,^..^,,,
> >
> >
> > On Fri, May 15, 2015 at 6:48 PM, Hristo Trendev <[email protected]>
> > wrote:
> > > Hello,
> > >
> > > I have installed the 2.0 developer preview and I cannot figure out how
> to
> > > compact a database. Trying with
> > >
> > > curl http://localhost:5984/<db>/_compact
> > >
> > > results in
> > >
> > > {"error":"forbidden","reason":"Compaction must be triggered on a
> > per-shard
> > > basis in CouchDB"}
> > >
> > > I have found this related GUI ticket
> > > https://issues.apache.org/jira/browse/COUCHDB-2492, but there is no
> hint
> > > how to actually start the compaction process manually. I have also read
> > in
> > > an email on the dev list, that this needs to be run against the 5986
> > port,
> > > but again no clue what the exact URL is. I have tried to guess it, but
> > none
> > > of the combinations that I've tried did work.
> > >
> > > Has anybody had any success with it?
> > >
> > > BR,
> > > Hristo
> >
>

Reply via email to