Using Riak 0.14.0-1_i386 on a 7 node cluster (mixed 686 & x86_64 Ubuntu and
CentOS Linuxes)

Filled a "person" bucket with more than 100.000 persons with simple "ID" ->
"NAME" document

Tried to query some data through HTTP API like that:

curl -X POST -H "content-type: application/json"
http://172.16.0.251:8098/mapred --data @findPerson.json
having the file findPerson.json like that:
{    "inputs":{
        "bucket":"persoane",
        "key_filters":[["starts_with","26"]]
    } ,
    "query":[
        {
            "map":{
                "language":"javascript",
                "source":"function(v) {return [v.values[[0]].data+' -
'+v.key];}"
            }
        },
        {
            "reduce":{
                "language":"javascript",
                "name":"Riak.reduceSort"
            }
        }
    ]
}

Everything worked ok, then I tried to query the database for counting
persons, changing the query with :
{    "inputs":{
        "bucket":"orase",
        "key_filters":[["starts_with","26"]]
    } ,
    "timeout": 900000,
    "query":[
        {
            "map":{
                "language":"javascript",
                "source":"function(v) {return [1];}"
            }
        },
        {
            "reduce":{
                "language":"javascript",
                "name":"Riak.reduceSum"
            }
        }
    ]
}

The problem I got is that the answer is a mixed result from answers computed
with the first JSON query and other computed with the second one.
It's just like a couple of servers within the cluster have done the
map/reduce functions USING the old javascript functions , remembering them
...

How could that be possible?
Constantin Teodorescu
_______________________________________________
riak-users mailing list
[email protected]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com

Reply via email to