2015-09-26 06:41 PM
We're running into an issue on one of our cDOT Cluster (8.2.3P2) with sporadic closed connections as indicated by the error below:
The workflow is a C# based program that uses our SDK and API calls on a W2K R2 based host.
While dedupe is running on a given Volume we check every ten seconds if dedupe is completed. The command succeeds most of the time but not always. If it fails we get this message:
Command was invoked on filer: 10.188.120.220
not authorized for that command ---> NetApp.Manage.NaApiFailedException: not authorized for that command at NetApp.Manage.NaServer.InvokeElem(NaElement element)
I was also able to repro the issue in a Powershell loop where I ran " Get-NcSis <vol-name>" 100 times. I got the error in iteration 26, 36, 39, 63, 67, 70, 75, 78, 79
I cannot find any errors in the logs about what looks like an RBAC issue on the Cluster itself (mgwd, debug, messages). I do see the incoming requests in the apache log. What could cause the Cluster to deny authorization for the API call seemingly on a random basis?
Any help is appreciated!
PS The cluster is not heavily loaded, so I doubt that load is a factor.