I tried testing the cluster with 'curl' as suggested on the Basho site:

 

curl -XPUT http://172.16.33.107:8098/riak/images/1.jpg -H "Content-type:
image/jpeg" --data-binary @004.jpg

 

I received no errors from this command it that means anything.

 

From: riak-users [mailto:[email protected]] On Behalf Of
Jeremiah Peschka
Sent: Tuesday, March 12, 2013 8:35 PM
To: riak-users
Subject: Re: Error interpretation

 

Default leveldb config is (when built using `make devrel`): 

 

 %% eLevelDB Config

 {eleveldb, [

             {data_root, "./data/leveldb"}

            ]},

 

Drop everything but the data_root. Also: simplify your set up. You're
running into timeout errors in a networked environment. Test your
functionality on a local cluster (either 1 node or a devrel setup with 4
nodes). Validate your logic and functionality there.

 

I wish I could promise that I can help out at all, but I just got back from
12 days out of the country and I'm booked solid for the next two weeks. In
short - my responses will be hit or miss at best.

 

Good luck and godspeed.




---

Jeremiah Peschka - Founder, Brent Ozar Unlimited

MCITP: SQL Server 2008, MVP

Cloudera Certified Developer for Apache Hadoop

 

On Tue, Mar 12, 2013 at 6:16 PM, Kevin Burton <[email protected]>
wrote:

The eleveldb app.config are the default. I don't know enough to modify them.
What are the defaults?

 

From: riak-users [mailto:[email protected]] On Behalf Of
Jeremiah Peschka
Sent: Tuesday, March 12, 2013 8:06 PM
To: riak-users
Subject: Re: Error interpretation

 

Your eleveldb app.config settings are decidedly non-default. Is there any
reason for that? Also, you spelled "cache_size" incorrectly in your eleveldb
config setting. What happens when you set those back to defaults? Is there a
reason that you're changing these settings?

 

This looks eerily similar to an unanswered Stack Overflow question [1] and
marginally similar to this mailing list thread about bad disks [2]. Since
it's after hours Pacific time, I'd wait for tomorrow and hope that some of
the folks living in "the future" have an idea.

 

[1]:
http://stackoverflow.com/questions/12748154/riak-riak-kv-vnode-worker-pool-c
rashed

[2]: http://www.mail-archive.com/[email protected]/msg09242.html




---

Jeremiah Peschka - Founder, Brent Ozar Unlimited

MCITP: SQL Server 2008, MVP

Cloudera Certified Developer for Apache Hadoop

 

On Tue, Mar 12, 2013 at 5:14 PM, Kevin Burton <[email protected]>
wrote:

On another node the console.log looks like:

 

2013-03-12 18:41:19.687 [info] <0.7.0> Application riak_kv started on node
'[email protected]'

2013-03-12 18:41:19.709 [info] <0.7.0> Application merge_index started on
node '[email protected]'

2013-03-12 18:41:19.719 [info] <0.7.0> Application riak_search started on
node '[email protected]'

2013-03-12 18:41:19.724 [info] <0.473.0>@riak_core:wait_for_application:419
Wait complete for application riak_kv (0 seconds)

2013-03-12 18:41:19.782 [info] <0.7.0> Application riak_api started on node
'[email protected]'

2013-03-12 18:41:19.846 [info] <0.7.0> Application cluster_info started on
node '[email protected]'

2013-03-12 18:41:19.867 [info] <0.7.0> Application riak_control started on
node '[email protected]'

2013-03-12 18:41:19.868 [info] <0.7.0> Application erlydtl started on node
'[email protected]'

2013-03-12 18:41:20.471 [info] <0.296.0>@riak_core:wait_for_service:439 Wait
complete for service riak_kv (1 seconds)

 

And in erlang.log  I see:

 

18:41:19.513 [info] New capability: {riak_kv,vnode_vclocks} = true^M

18:41:19.526 [info] New capability: {riak_kv,legacy_keylisting} = false^M

18:41:19.541 [info] New capability: {riak_kv,listkeys_backpressure} = true^M

18:41:19.558 [info] New capability: {riak_kv,mapred_system} = pipe^M

18:41:19.582 [info] New capability: {riak_kv,mapred_2i_pipe} = true^M

18:41:19.623 [info] Waiting for application riak_kv to start (0 seconds).^M

18:41:19.687 [info] Application riak_kv started on node
'[email protected]'^M

18:41:19.709 [info] Application merge_index started on node
'[email protected]' ^H^M

18:41:19.719 [info] Application riak_search started on node
'[email protected]' ^H^M

18:41:19.724 [info] Wait complete for application riak_kv (0 seconds)^M

18:41:19.782 [info] Application riak_api started on node
'[email protected]'^M

18:41:19.846 [info] Application cluster_info started on node
'[email protected]'^M

18:41:19.867 [info] Application riak_control started on node
'[email protected]'^M

18:41:19.868 [info] Application erlydtl started on node
'[email protected]'^M

Eshell V5.9.1  (abort with ^G)^M

([email protected])1> 18:41:20.471 [info] Wait complete for service riak_kv
(1 seconds)^M

 

===== ALIVE Tue Mar 12 19:09:45 CDT 2013

I don't see any errors there.

 

There is an empty crash.log that was dated with a timestamp recently.

 

From: riak-users [mailto:[email protected]] On Behalf Of
Jeremiah Peschka
Sent: Tuesday, March 12, 2013 6:34 PM
To: <[email protected]>
Subject: Re: Error interpretation

 

Your Riak search hook crashed. There is a bad argument being sent to Erlang.
Check your Riak logs for more detail.


--

Jeremiah Peschka - Founder, Brent Ozar Unlimited

MCITP: SQL Server 2008, MVP

Cloudera Certified Developer for Apache Hadoop


On Mar 12, 2013, at 4:23 PM, "Kevin Burton" <[email protected]>
wrote:

I am writing a value to a 4 node Riak cluster with a CorrugatedIron client
and am getting the following error:

 

Riak returned an error. Code '0'. Message:
{precommit_fail,{hook_crashed,{riak_search_kv_hook,precommit,error,badarg}}}

 

I don't know how to interpret this error. Any help would be appreciated.

 

Thank you.

_______________________________________________
riak-users mailing list
[email protected]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com

 

 

_______________________________________________
riak-users mailing list
[email protected]
http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com

Reply via email to