Customer is receiving a hub alarm for too many subscribers.
How to decrease the number of hub subscribers when you receive a hub alarm for 'too many active subscribers'
In general, we recommend that you switch to using a Linux hub or try to cut down on the number of hub subscribers on the current Primary hub (Windows OS). Note that 'temp' subscribers also count in the total number of subscribers. temp subscribers can be IM GUI sessions as well.
Note that there are 2 kinds of subscribers:
1. Queues, which are configured explicitly on the hub or other hubs.
2. Temporary subscribers, which listen for certain messages but do not have defined 'queues'
- Use a Linux versus a Windows Hub because the hub can only handle a maximum of 64 subscribers in a Windows environment due to Windows OS (winsock) limitations on the number of open sockets.
- Add a new/additional hub layer to your UIM architecture
You may need to add another layer to your UIM architecture, e.g., where your Primary (CORE) hub connects to multiple intermediary hubs (also in the central location), which in turn connect to the remote/satellite hubs and/or then customer/client hubs.
- When possible move GET queues to a middle tier/Tier 2 hubs/tunnels, and set up an associated single ATTACH and GET queue on the primary to receive the alarms/QOS from each of the downstream hub(s), e.g., 3-4 hubs in the middle tier/layer.
- IM sessions: Tell UIM administrative users to log off of IM on the Primary hub and use middle tier hubs/tunnel servers when they need to access and use IM. Also admin users could use the Admin Console instead of IM sessions
- Make sure DrNimBUS is not running because it creates a busy temp queue
- Delete the baseline_engine/prediction_engine and associated queues from the Primary hub
Besides asking your admin users to login to other hubs versus the Primary, in general on a primary, you could also deactivate the following probes on the primary hub, e.g., if youre not using the audit feature or the fault correlation.
baseline and prediction should be running on remote hubs, not the primary since the primary should be dedicated to core fuinctions, as opposed to monitoring tasks.