Virual Node Abnormality

Mark Phillips mark at basho.com
Fri Nov 4 12:01:18 EDT 2011


Hi Vicky,

Sorry for the late response here.

Bad news, good news, and better news ..

Bad - looks like you're hitting this issue bug -
https://issues.basho.com/show_bug.cgi?id=1250
Good - it was fixed in 1.0.1 [a] (and perhaps you've already upgraded)
Better - 1.0.2 will be out *very* soon (we are in the final testing
stages). You're probably best to hold tight for a bit and upgrade from
1.0.0 -> 1.0.2

Mark

a - http://lists.basho.com/pipermail/riak-users_lists.basho.com/2011-October/006121.html




On Thu, Oct 13, 2011 at 7:36 AM, Vicky Khanna
<vicky.khanna at position2.com> wrote:
> Hi,
>
>
>
> I am testing 1024 ring size 4 node riak system.
>
> I am facing a virtual Node abnormality in which only one node has the
> following error while performing handsoff
>
>
>
> =ERROR REPORT==== 13-Oct-2011::14:10:26 ===
>
> Failed to merge
> ["/test/riak/bitcask/689360635576978622551152123244106150872866619392",
>
>
>
> This is creating two file every 3 mins in the same folder. The file which is
> created has a size 0 for both .data as well as for .hint.
>
> These files which get accumulated every 3 mins are increasing the total no
> of open files by riak.
>
> So I have to unnecessary increase the ulimit to a very big no.
>
>
>
> ls -l of the riak data directory gives the size of abnormal virtual node to
> be 954368 (4096*233 ) where as all 1023 virtual nodes have a folder size of
> 4096.
>
>
>
> Please tell me if I can solve this problem by any means and if I have done
> something wrong in configuring the system.
>
> Also if I m bound to have any data loss…???
>
>
>
> Thanks and Regards,
>
>
>
> Vicky Khanna
>
>
>
> _______________________________________________
> riak-users mailing list
> riak-users at lists.basho.com
> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
>
>




More information about the riak-users mailing list