An essential part of Array management is monitoring for specific events or error conditions. When an event or error does occur Unisphere displays an alert and, if configured to do so, notifies you of the alert by the way of Email, SNMP or Syslog. If you wish to find out more about configuring SNMP and the Syslog please refer to this guide.
For a brief over view please click here.
To begin with let’s start with the Alert Policies section. In the following examples we have focused on a 5977 PowerMax array. To navigate to here please select the Settings button as displayed by a “cog” icon in the upper right hand corner of Unisphere.
When you install Unisphere some of these categories come enabled by default. However in order to capture a complete picture of the health, capacity and behaviour of the array we would recommend the following categories you can see to be enabled:
- Array Component Events
- Array Events
- Device Config Change
- Director Status
- Hotspare Invoked
- Port Link Status
- Port Status
- SRDF Link Status
- Storage Pool Projection Alert
You can see here that the notifications turned on is email. It can also be via SNMP or the syslog.
To configure these please select the Notifications on the left hand side navigation:
Here select the configure button below email in order to set it up. You can configure email addresses to which notifications, alerts, and reports are sent. You can configure a single email address for all notification instances, or you can use different email addresses for different notifications on different storage systems.
In the Outgoing Mail Server (SMTP) section you need to specify the IP Address or Host and also the port you are using. In the User Information you will need to provide a sender email address that will make sense and here we have just called it out as Array0111@dell.com. You select create to configure an email address that will receive the alerts, in general it will be the storage team monitoring this so firstname.lastname@example.org would be a valid and relevant mail address.
Finally select which array and category which is relevant to you so here we have selected System, Performance and reports as we want to be alerted on these.
Compliance Alert Policies
This section is concerned with monitoring the service level compliance state of our storage groups.
Here we have selected all the storage groups on array 0111 and the state we wish to monitor for is marginal or critical as we are not concerned with a stable SG state.
As you can see here we can have enabled email notifications for a certain number of SG’s if their state goes marginal or critical.
Performance Threshold and Alerts
This category will allow you to monitor the overall performance of the array and alert on it accordingly.
In this example we are focused on the 0111 Array. Here we are focused on the array category which will allow us to track and alert when we go over a certain number in terms of a threshold across a variety of categories that we can see on the list. You can see that this list includes the KPI’s and generally the KPI’s that we pre-configure for the customer will satisfy most people’s needs but you may want to adjust the values based on your own preference.
As well as Array there are 5 other categories such as Storage & Hosts, Directors & Ports, Storage Resources, Data Protection and System Internals.
So I wanted to show a use case example here so I will select Storage & Hosts > Storage Groups.
So for this example I am very sensitive about the Response Time of my SG’s on this array so I choose the pencil icon next to RT so I can edit it.
The default values were 20 and 30 so I have reduced them to 10 and 20. Next I select the trigger alert on the lower part of the screen.
I have selected the checkbox for trigger an alert and my first threshold is just a warning but the 2nd is a critical alert. The number of occurrences in the data samples which must happen before the alert is triggered. For example, if the threshold is breached 3 times out of 5 samples, an alert is imitated. The number of occurrences in the data samples which must happen before the alert is triggered. For example, if the threshold is breached 3 times out of 5 samples, an alert is initiated.
As I stated previously most people are happy to accept the pre-configured thresholds that are automatically setup when you install Unisphere. If you do wish to change these you have that ability and we have shown you a practical example above.
Symmetrix Threshold Alerts
These are set of alerts based on certain categories such as capacity consumption, Meta Data Utilization and Local Replication Utilization.
In this example the array is a 0111 PowerMax box so we have 6 categories Storage Resource Pool Utilization, Storage Container Utilization, Frontend Meta Data Usage, Backend Meta Data Usage, Local Replication Utilization and System Meta Data Utilization. There are different categories for older arrays such as Fast and thin pools but are not applicable to the newer arrays.
In terms of the 3 various thresholds Warning, Critical and Fatal we have gone with the default values. We can of course customize these if we wish also and in terms of notifications we have again chose email but SNMP and syslog are also available.
Following these guidelines will enable you to successfully monitor your array's performance. I hope you have found this blog helpful and if you have any questions please feel free to reach out with questions.