couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Newson <robert.new...@gmail.com>
Subject Re: CouchDB for logging & file storage
Date Mon, 22 Jun 2009 17:12:24 GMT
yes, stale=ok won't help if you've never built the view. but if you
hit your view without stale=ok every N minutes from a cron job or
something, then you can always use stale=ok from your "real" clients,
and they'll never block for a view update. You'll always be five
minutes behind, but I think that's probably ok.

B.

On Mon, Jun 22, 2009 at 4:36 PM, Sergey Shepelev<temotor@gmail.com> wrote:
> Yeah, but for new queries it would return empty set until all documents are
> processed, right?
> This makes debugging new queries a bit tricky.
>
> You have to pick unbiased 2-3K documents into separate database, test your
> queries there while they're fast to build and then use them on big database.
>
> On Mon, Jun 22, 2009 at 7:29 PM, Robert Newson <robert.newson@gmail.com>wrote:
>
>> You can also pass stale=ok if you don't need the latest results from a
>> view.
>>
>> http://wiki.apache.org/couchdb/HTTP_view_API
>>
>> B.
>>
>> On Mon, Jun 22, 2009 at 3:46 PM, Sergey Shepelev<temotor@gmail.com> wrote:
>> > From my poor experience with couch, i think that your task (ever
>> expanding
>> > set of data, no updates) is perfect for couch.
>> > But, you should
>> > also consider that expanding-only set of data is a good task for
>> > constant databases too. Though, you would run into a expansive process
>> > writing a nice querying layer on top of key-value CDB.
>> >
>> > You can do basic and somewhat advanced querying with couch, though, full
>> > text searching is not here. (Again, my knowledge is poor, maybe FTS is
>> > possible)
>> > Consider also, that couch indexing is lazy, it doesn't index documents
>> until
>> > you ask it a query.
>> >
>> > AND, the most difficult (for me) part - query is blocking on indexing.
>> You
>> > won't have any results until whole database is indexed. Index results are
>> > cached, so old documents aren't processed until they changed. But for new
>> > type of query and database with 300K items you will wait tens minutes to
>> > hours to get query result.
>> >
>> > But you can run a tiny daemon that does query few times a minute so you
>> > always have fast queries with up-to-date data.
>> >
>> > On Mon, Jun 22, 2009 at 6:01 PM, Nitin Gupta <
>> nitin.gupta@srishtitechnet.com
>> >> wrote:
>> >
>> >> Hi All,
>> >>
>> >>
>> >>
>> >> I am new to non-rdbms storage systems. I am working on an application in
>> >> which we need to log different user actions as plain text. I am not of
>> the
>> >> opinion to make use of our rdbms for this activity as writing this on a
>> >> rdbms can be an expansive process.
>> >>
>> >>
>> >>
>> >> Can I make use of Couch DB for this activity? Also I am looking to store
>> >> all
>> >> the attachments or BLOB into some non-rdbms solution. Can I make use of
>> >> CouchDB here as well. This way I will just have to add one network
>> >> component
>> >> to address file storage and logging of user action.
>> >>
>> >>
>> >>
>> >> I would probably need to show the logged information on the admin
>> console
>> >> or
>> >> some other UI to the users. I hope that I can do basic indexing,
>> querying
>> >> etc on the basis of data items.
>> >>
>> >>
>> >>
>> >> Regards,
>> >>
>> >> nitin
>> >>
>> >>
>> >
>>
>

Mime
View raw message