Using IBM Storage Insights Pro and Alert Policies To Monitor Host Path Count

I was at a recent TechU event and had a discussion with a customer about using IBM Storage Insights to monitor host path count.  More specifically, the customer had a recent outage of a few hosts after doing some maintenance work on some storage and the affected hosts were not connected on all the expected paths.    There are options for using the storage data collections to review host connections.

For instance, you could write a script that compares the SVC/Storwize host WWPN definitions to the connected devices to see if any WWPNs are not connected to the SVC.  However, it is much more straightforward to configure an alert in IBM Storage Insights Pro.   Alert Policies were previously covered in these two videos: 

You can create an Alert Policy for an agentless host and then configure an alert to notify you when the path count changes.  You would need to create separate policies for each operating system, but after you create the first policy and define the alert, you can copy that policy and modify it for each operating systems.  In the following screenshot, I am creating a policy for Windows servers and I am adding all of my Windows servers to this policy.

In the next screenshot, I am adding an alert to the new policy. The basics of adding alerts are covered in the videos above.    You can see that I have selected a resource type of "Disks" and then "General".   Under General, I check the box next to "Paths" and then in the Condition drop-down, I select Changes.  I am interested in path count changes.  Alternatively, you could set a value for your expected path count and alert when that falls below some value, but different servers may have different expected path counts. 

Next,  I configure the email address to send the alerts to.   This may be a group email address for your team, or perhaps it is one of your host administrators.  I also set the notification frequency for once for every violation, although you could set it to notify you some number of minutes, hours or days.  I have also set the Alert level to the Warning, rather than Critical, but your requirements may be different. 

After you set those options, save the new alert.  Alerts will now be generated when the path count changes for the hosts included in your new Alert Policy.  You can remove hosts from the policy if you do not want alerts generated for them.  Perhaps you only want to monitor your more critical hosts. 


Popular posts from this blog

Troubleshooting Slow Drain Devices on Broadcom Switches

Spectrum Virtualize NPIV and Host Connectivity