Launch a Cube!

[danigonzalem]

Hello everybody,
I hope someone answer me, because I never get the complete solution. I want to launch a cube using the API and when the cube is executed get the status in order to relaunch if it has been an error or notify to the users when apply.

I used this worflow:

1. Get the datetime of the last update
2. Launch the cube
3. When the cube has the status 64 (finished)
   • if updated datetime<>initial datetime then OK
   • if updated datetime=initial datetime then KO

I Want to know if this is the correct way to ensure that a cube is launched correctly, thanks!

[xdumstr]

You can use monitor cache API to check the cube cache status.

[danigonzalem]

Thank you xdumstr,
we do this for checking the lastupdatetime of the cube's cache
GET $url_base+"/MicroStrategyLibrary/api/monitors/caches/cubes?clusterNode="+$clusterNameNode+"&projectIds="+$projectId
with this we get a JSON where there are all cubes and their caches, but not always is updated this field. We check the JSON for 15 minutes after that the job returns a KO.
Is it possible? should we do the call "GET /api/monitors/caches/cubes/{cacheId}" better??

Bye

[elinmstr]

@danigonzalem The lastUpdateTime field is the last time the cube was compared to the data warehouse. To have this updated through rest api, you must execute/publish the cube to library each time POST /api/v2/cubes/{cubeId}.
In summary, this workflow

1. GET /api/monitors/caches/cubes/{id}
2. POST /api/v2/cubes/{cubeId}.
3. GET /api/monitors/caches/cubes/{id}

should work. In step 2, a new job is created each time which will update the cache after it's done.

From step 1 to step 3. The json field lastUpdateJob would be updated along with lastUpdateTime for instance if the lastUpdateJob in step 1 was 100 in step 3 it would be 101 assuming no other jobs were running.

Note: Make sure the cube's state is correct. For instance, is it active and loaded in step 1 and step 3(loaded into memory).

[danigonzalem]

Hello and thanks!
My objective is avoid to set the node for looking up the caches, that is needed in /api/monitors/caches/cubes and in the /api/monitors/caches/cubes/ is needed the id of the cache of the cube, not the definition of the cube. An the id cache of the cube is in /api/monitors/caches/cubes too.

So finally we wil use this flow:
1.GET /api/cubes/{id} -> get lastupdatetime
2.POST /api/v2/cubes/{cubeId} -> launch cube
3.HEAD /api/cubes/{cubeId} -> verify the status
4.GET /api/cubes/{id} get lastupdatetime to confirm the new update date

Thanks for helping me