Mike | 28 Feb 00:46 2015

couch to influx

Hi Everyone,

I have started another couch stream follower - this time for influxdb,
it is more of a test project for me to mess with influxdb and grafana but
could be of use for someone else.

Its not very complete but will follow a couchdb stream and insert 
counters and
gauges into an influxdb.  I am not sure on the correct data format just 
yet so any
tips pointers most welcome.



Vladimir Ralev | 27 Feb 12:46 2015

Are views replicated or built separately on each cluster node?


Need to ask if the view indexes or view files or may be just chunks of
views would be copied between cluster nodes or only the documents are
copied and each node maintains its own index independently from each

jumbo jim | 25 Feb 23:37 2015

Atomic transactions, race conditions

I was inspired to some atomicity testing with transactions and couchdb, and
testing under race conditions.

I from this blog entry -


Firstly, I am having problems getting the concept working and want to be
sure I understand the nature of views correctly.

Firstly, I am working with a single node. So, no possibility for
replication delay issues etc.

Secondly, I understand there can be 3 options when working with views -
default, stales=ok, update-after.

I have not set "stales=OK", or "update after", therefore I assume I am
working with the default behaviour of views.

I am interested with the "oldest pending" view example as shown in the blog
- in particular the reduce portion. That view will show the transaction
with the oldest timestamp.

I understand that the default behaviour when I call a view - the view will
recalculate *at that point in time*, and may even take a while depending on
the amount of documents involved. There should be no risk of stales?

I am testing against race conditions. I wrote a small piece of code to
insert a document which includes an amount and timestamp, I then
immediately call the view. When I say immediately - it could be within 1ms
(Continue reading)

Mike | 25 Feb 21:05 2015

couch wiki - utils page

Hi All,

I have added a couple of bits to the 
page and realised it is a little jumbled.

Would it be sensible to split it up into clearer sections eg:

Couch Apps
    Futon Chrome Bookmark App
    couchdb-utils - A fast and portable CouchDB command line utility (Go).
    couchapp - popular command line utility for managing design 
documents (Python).

Search/External Applications
    Elasticasearch -    River info + links to tutorial ?
    Lucene -CouchDB-Lucene <https://github.com/rnewson/couchdb-lucene> - 
A full-text indexer...

     munin-plugin-couchdb - Munin plugin for graphing CouchDB statistics.
     statsd backend - Backend for StatsD to emit metrics to CouchDB.

    jq - A command-line JSON processor. You can use it to parse, filter 
and manipulate CouchDB's JSON output.
    jgrep - A Ruby-based CLI tool and API for parsing and displaying 
JSON data using a logical expression syntax.
(Continue reading)

Andrew Koster | 25 Feb 14:43 2015

How to query a view from within another view?

Dear all,

I have a CouchDB database with the following type of documents, representing
events that happen:

{ person: 1

  timestamp: 1

  event: { type: enter

           room: b }



{ person: 2

  timestamp: 5

  event: { type: leave

           room: b }


Now the problem that I want to solve is the following: I want to know the
total amount of time that every other person spent in the same room as
person 1, at the end of the day when all persons have left all rooms. Note
(Continue reading)

Michael Power | 24 Feb 22:41 2015

Migrating couchdb data from one version to another

When we are ready to put our new application on production we want to make sure the couchdb is ready to receive
it.  For example the app may need new views, we may want to migrate some data, we may want to bulk insert some
data.  Do you guys use any external tools to maintain the state of your couchdb database?  I was looking for a
couchdb equivalent for mybatis migrations http://mybatis.github.io/migrations/.

⌘Michael Power
michael.power <at> elotouch.com<mailto:michael.power <at> elotouch.com>

Lena Reinhard | 24 Feb 15:18 2015
Robert Kowalski | 23 Feb 09:33 2015

How to organize an event like CouchDB Day

Hi list,

CouchDB Day was quite fun and Andy and me hope for more CouchDB Days in the whole world, so I wrote a short introduction how to organize events like the CouchDB Day: http://robert-kowalski.de/blog/how-to-organize-an-event/

Lena Reinhard | 17 Feb 21:24 2015
Roald de Vries | 16 Feb 16:58 2015

Stop replication immediately

Hi all,

I have a problem with stopping replication.

I have 1 main database containing all data, and then per-user databases containing limited data. The
user-databases are filled by filtered continuous replication. Specifically, some user-metadata
document for the user is replicated, containing a field for `receivesUpdates`. Now if I turn that field to
`false`, I want the replication job to stop immediately, and not send a document anymore. It could be a
privacy issue if another document is replicated after setting `receivesUpdates` to `false`.

The best I’ve come up with so far is to have a changes-feed listener that stops the replication job when it
detects such a change in the main database. But before the replication is stopped, I guess another change
might have come through.

My question is: is there a solution for this problem? And if not: is that fundamental? Is the changes feed
still consistently ordered in a cluster?

Thanks in advance, Roald
Ben Ng | 11 Feb 10:14 2015

Why are temp views so much faster than design docs?

Hi all,

I’m perplexed. Using the exact same map function, and no reduce function, my temp views are performing
significantly faster than my design doc during the initial indexing phase.

The temp view finishes loading in 26 seconds, while the design doc is on track to take 2-3 hours.

I am using Couch 1.6.1.

What could be the problem here?