Jeremy,
looks like you're hitting the timeout for the MapReduce job you're running, so
the issue isn't memory-related. You could either increase the timeout for the
whole job, or split up the single MapReduce request in multiple, ensuring a
shorter runtime of a single job, then collecting the results on your
application's side.
Mathias Meyer
Developer Advocate, Basho Technologies
On Dienstag, 21. Juni 2011 at 14:25, Jeremy Raymond wrote:
> I increased the memory to 3GB on the VMs I'm using for Riak and also replaced
> a JavaScript reduce function I had missed converting to Erlang with the
> Erlang version. Monitoring the memory on the machines indicates that Riak is
> not running out of memory. There is lots of disk space on the machines (~30GB
> free).
>
> Riak crashed again this morning with a {error, flow_timeout} (below). My use
> case is I have ~75 map/reduce jobs that run about every 10 minutes over about
> 8700 items in a bucket. Once an hour the 8700 items get updated with a write
> to each item.
>
> =SUPERVISOR REPORT==== 21-Jun-2011::02:34:34 ===
> Supervisor: {local,luke_phase_sup}
> Context: child_terminated
> Reason: {error,flow_timeout}
> Offender: [{pid,<0.7243.26>},{name,undefined},{mfa,{luke_phase,start_lin\
> k,[riak_kv_reduce_phase,3,[accumulate,converge],undefined,<0.7241.26>,66000,[{e\
> rlang,{reduce, ...
>
>
> =ERROR REPORT==== 21-Jun-2011::02:34:34 ===
> ** State machine <0.7274.26> terminating
> ** Last message in was {'EXIT',<0.7241.26>,{error,flow_timeout}}
> ** When State == waiting_kl
> ** Data == {state,<0.7241.26>,mapred,<<>>,
> [[{639406966332270026714112114313373821099470487552,
> '[email protected] (mailto:[email protected])'},
_______________________________________________
riak-users mailing list
[email protected]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
