Riak high IO usage

Magnus Kessler mkessler at basho.com
Sat Jul 16 04:25:53 EDT 2016


Hi "!"

Could you please sign with a name, so that we know how to address you?


On 15 July 2016 at 07:45, ! <linwaterbin at qq.com> wrote:

> hi all,
>   there is a program comfusing me that Riak produce continuous high IO
>

[graph showing IO spikes reaching 100% for minutes at a time removed]

Unfortunately you don't give us a lot of information to work with. The
graphs show spikes in IO reaching 100%, but I would not call that
continuous, unless you mean that whenever there is traffic it reaches 100%
IO and stays that way until traffic drops again.

It is quite possible to max out the available IO bandwidth when PUTting or
GETting large amounts of data from Riak. What workloads are you running
against your Riak installation? How are your storage volumes configured?

We always recommend to use the fastest storage solutions you can afford,
and to run tests with workloads that are as similar as possible to the
expected production workload of your application.



>   when IO coming,I  observer a phenomenon :
>

[image of console showing file tmp_agent_at5SIa removed]

   what is tmp_agent_XXXX ? what is the relationship between it and the
> high IO ?
>

There is no mentioning of tmp_agent anywhere in Riak's code base. Please
check your OS installation for possible causes of this file appearing. You
may want to remove the file from /var/lib/riak, as it has no place there.

Please give us some more information, if you'd like to pursue this
investigation further.

Kind Regards,

Magnus


-- 
Magnus Kessler
Client Services Engineer
Basho Technologies Limited

Registered Office - 8 Lincoln’s Inn Fields London WC2A 3BP Reg 07970431
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.basho.com/pipermail/riak-users_lists.basho.com/attachments/20160716/b9c10623/attachment-0002.html>


More information about the riak-users mailing list