Hi. We recently installed OnCommand Clustered Data ONTAP 126.96.36.19974 management pack into our SCOM and it really slowed our Ops Console. Right click on any alert freeze it for approx. 20 sec every time.
I believe it is because this MP created over 400 internal connectors (Administration - Product Connectors - Internal Connectors) which look like this.
340 of them:
|OC.Cluster.OM.ClassType.Connector.0065426f-9a81-4176-96d3-ee1c7a2edfac||Clustered Data ONTAP MP uses this connector to save discovered cluster storage components to SCOM database|
65 of them:
|OC.Cluster.OM.Connector.fd30feca-0613-11e6-b3ef-00a098a0623f||Clustered Data ONTAP MP uses this connector to save discovered storage virtual machine objects to SCOM database|
|DataONTAP.Cluster.ClusterVserverSCOMConnector||Clustered Data ONTAP MP uses this connector to save ClusterVserverSCOMConnectors data.|
|DataONTAP.Cluster.DashboardConnector||Clustered Data ONTAP MP uses this connector to save dashboard data to SCOM database.|
|DataONTAP.Cluster.MgmtServerDiscoveryConnector||Clustered Data ONTAP MP uses this connector to save discovered management servers to SCOM database.|
|DataONTAP.Cluster.StorageDiscoveryConnector||Clustered Data ONTAP MP uses this connector to save discovered cluster storage to SCOM database.|
Anyone of you noticed something similiar and if it so, would it be okay to remove those connectors ?
We have also different NetApp which we monitor with OnCommand Data ONTAP 188.8.131.5200 and there we have only 5 connectors established.
We too are experiencing this issue. We have uninstalled the OnCommand plugin and it has left the connectors. Within our Active Alerts screen we are getting delays of up to 10 seconds before we can close an alert.
We have approximately 200 connectors that have been added.
Please can someone provide a fix? It is not supported by Microsoft to remove these connectors.
I raised case with NetApp and this is what I received as answer, (but this is not acceptable for us).
"In our software we create a connector per storage object to make parallel operation possible. Therefore, also with fewer controllers and high count of storage objects you could reach a high number of connectors.
What could be done?
If you have more servers, you could spread the workload by adding servers to the pool and to add some controllers to other management servers. If you have more than 1 they should appear under management servers."
The issue with slow alert view and right click is because "Forward to" option which search through all the connectors.
is this case still open with NetApp? @Fursel
Did you find a workaround?
thanks for a feedback.
kind regards from far away
We was able to resolve this usse, once the management pack was removed the connectors still remained which made the SCOM console really unresponsive.
The fix was to remove the connectors via powershell. Use Remove-SCOMConnector. Be careful when using this and make sure you read the documentation.
thanks for the Feedback @James_Goodison
Are you still using the NetApp MP? If so - how many connectors do you currently have? If you are not using the MP anymore - have you found an alternative?
kind regards from far away
You can use the below link to remove the old connectors
Thanks Kevin for the Link
Hi NetApp Team.
I have observed that Netapp connectors are rebuilding once we run the discovery, looks like Net App MP inbuilt the connectors in the discovery code.This is happening only on Clustered Data ONTAP not on 7-mode Data ONTAP MP.
Due to these connnectors, Console is not doable at all.
Please look into this issue as many customers are facing with same issue and request to remove these connectors from the discovery code,
or let us know how we can remove these connectors permanently.
No sorry we didnt find solution for that. What I can think of is to add another management servers and divide monitoring of those storages to them like NetApp said or just turn off automatic discovery and do it manually once per change with forcing script afterwards to delete those connectors.
Thanks Fursel for the updates.
This is something really looked into the next version, adding a new server approach needs to be tested and am not sure that loading the connectors per alert patern would change either , I tried the other solution, when we ran maually it recreate all the connectors again.
We are experiencing the same problem. Tried to delete the connectors with remove-scomconnector command, it gave some fails, then tried to fix it with powershell script, sql script. It seemed working well, but when I tried to approve manual installed agent, it threw out one exception. So I must restore the database and scom management servers from backup. Could Netapp help to fix this as soon as possible ?
Thanks in advance.
Looks like no updates from NetApp end. I have tried to load balance the MP with multiple Management Servers.. still no improvements in performance..
Please let us know if any solutions would help us to improve the performance
NetApp, please consider another approach for this. Contextual menus when dealing with Alerts dashboards are becoming increasingly slow to render due to the huge quantity of connectors being loaded for the "Forward to" menu when right clicking on alerts.
More than a year past by now. I don't think NetApp will take any action.
Please NetApp, could you at least give some information about this? Even if you decide to do nothing at all!
678 connectors, freeze for ca. 50 seconds