I think this may be the bug that is causing the problem:
   https://issues.apache.org/jira/browse/USERGRID-1270

As I understand it, that problem has to do with excessive tombstones
created by Graph Edge sharding.

Two things you can try:

1) Create a build from the release-2.1.1 branch where this bug was fixed,
and give it a try

And / or:

2) Flush the Cassandra tables that are related to sharding, which should
clean-up tombstones.

The graph-related tables are:

   Graph_Source_Edge_Types
   Graph_Source_Edge_Id_Types
   Graph_Target_Edge_Types
   Graph_Target_Edge_Id_Types

   Graph_Source_Edge_Types_V2
   Graph_Source_Edge_Id_Types_V2
   Graph_Target_Edge_Types_V2
   Graph_Target_Edge_Id_Types_V2

   Graph_Marked_Nodes
   Edge_Shards

   Graph_Source_Node_Edges
   Graph_Target_Node_Edges
   Graph_Source_Node_Target_Type
   Graph_Target_Node_Source_Type
   Graph_Edge_Versions

Hope that helps...

Dave

On Sun, Apr 3, 2016 at 1:20 AM Thành Vũ Trung <[email protected]>
wrote:

> Hi all,
> When I execute DELETE method:
> http://localhost:8080/myorg/myapp/artices?ql=select *&limit=1000
> Tomcat7 used to about 90% CPU and any other request could not response. It
> keep that a long time and response server error. I know currently usergrid
> is not support to delete a large entities but if have a hacker try to
> execute above request, our system will be affected.
> Have any solutions to solve this issue?
>
> --
> Regards,
> Thanh.
>
>

Reply via email to