Percolate match error


(kj) #1

Hi,

The test case we are trying requires registering multiple percolate queries
and then using percolate API to determine matches.

The steps are as follows:

  1. Delete existing concrete index.
  2. Re-create concrete index
  3. Register multiple percolate interest/queries using the concrete index
  4. Perform match

The registration seems to go well but the match fails with the following
error.

{"error":"NullPointerException[null]","status":500}

This seems to happen only if we register multiple queries after deleting
and recreating the concrete index. It does not appear to be problem with
single percolate query. The error also goes away if the same queries are
re-registered and a delay is introduced between creation of concrete index,
registration and actually matching. Though the delay is not consistent, the
match then happens successfully with no errors.

Is there a way to avoid this error?

Thanks for the help,
KJ


(kj) #2

Any suggestions?


(Jörg Prante) #3

Explicit refreshing all indexes between registering the queries should do
instead of the delay, e.g. curl -XPOST 'http://localhost:9200/_refresh'

On Thursday, July 5, 2012 5:06:56 PM UTC+2, kj wrote:

Any suggestions?

--
View this message in context:
http://elasticsearch-users.115913.n3.nabble.com/Percolate-match-error-tp4020102p4020128.html
Sent from the ElasticSearch Users mailing list archive at Nabble.com.


(kj) #4

Thanks but the error still occurs.


(Clinton Gormley) #5

This seems to happen only if we register multiple queries after
deleting and recreating the concrete index. It does not appear to be
problem with single percolate query. The error also goes away if the
same queries are re-registered and a delay is introduced between
creation of concrete index, registration and actually matching. Though
the delay is not consistent, the match then happens successfully with
no errors.

Is there a way to avoid this error?

Please open an issue with a testable recreation of the error - that way
it'll be easier to confirm that it is fixed

thanks

Clint


(system) #6