Dell Unity: How to Enable CPU and Memory Utilization Monitoring (User Correctable)
Summary: This article describes how to enable CPU and Memory Utilization Monitoring, set custom system usage threshold alerts, and define user email messages to be notified.
Instructions
Note: As of Unity OE version 5.4, 5.5, this feature is not working correctly. Dell is investigating the cause of the issue; watch this Dell Knowledge Base article for further updates.
A workaround is available. Contact Dell Technical Support or your Authorized Service Provider and quote this Dell Knowledge Base article ID.
This feature simplifies monitoring array CPU and memory usage which otherwise must be done manually, while logged on to the system to view the system activity.
CPU and memory utilization on both array storage processors can be monitored and have alert notifications that are sent to users once user-defined thresholds are reached.
Initial high usage alerts are monitored using the sys_monitor job, within the UDoctor service on the array.
Also, resolve alerts are monitored using a separate sys_monitor_resolve job, also within the UDoctor service. The resolve alerts automatically clear the initial high usage alerts on the Unisphere UI once CPU or Memory usage has retreated below a separate user-defined threshold.
To enable the monitoring feature
-
Log in to the primary SP
-
Enable the sys_monitor and sys_monitor_resolve jobs using the UDoctor service wrapper script by running the following commands, respectively:
svc_udoctor -j -e sys_monitor svc_udoctor -j -e sys_monitor_resolve
To modify the run frequency of each job
-
Log in to the primary SP
-
Enter the UDoctor service wrapper interface for either the sys_monitor or sys_monitor_resolve job by running either of the following commands, respectively:
svc_udoctor -j -m sys_monitor svc_udoctor -j -m sys_monitor_resolve
-
In the following menu interface, select the SCHEDULE_TYPE option by typing 3 into the interface and press the return key:
Choose job property you want to modify 1) ENABLE 3) SCHEDULE_TYPE 5) GROUP 7) SAVE|QUIT 2) COMMAND 4) SCHEDULE_TIME 6) CLOUD_DEPENDENT #?
-
Choose your scheduling type by typing your selection of the following (7) options into the interface and press the return key:
Seconds|Minutes|Hours|DailyFixed|Days|Weeks|Months
-
Select the SCHEDULE_TIME option by typing 4 into the interface and press the return key.
-
Input your scheduling frequency by typing a numerical value represented by your unit of measurement, from step 4, into the interface and press the return key.
-
Select the SAVE|QUIT option by typing 7 into the interface and press the return key to save your changes and quit the UDoctor interface.
-
Upon exiting the interface, verify that the displayed configuration output matches your changes:
[sys_monitor] ENABLE = yes COMMAND = perl /opt/UDoctor/FE/bin/sys_monitor.pl -mon -cpu 90 -mem 90 SCHEDULE_TYPE = Minutes SCHEDULE_TIME = 15 GROUP = monitoring CLOUD_DEPENDENT = no
To modify the alert thresholds of each job:
-
Log in to the array’s primary SP.
-
Enter the UDoctor service wrapper interface for either the sys_monitor or sys_monitor_resolve job by running either the following commands, respectively:
svc_udoctor -j -m sys_monitor svc_udoctor -j -m sys_monitor_resolve
-
In the following menu interface, select the COMMAND option by typing 2 into the interface and press the return key:
Choose job property you want to modify 1) ENABLE 3) SCHEDULE_TYPE 5) GROUP 7) SAVE|QUIT 2) COMMAND 4) SCHEDULE_TIME 6) CLOUD_DEPENDENT #?
-
Update the job’s command with your thresholds by reentering the full command with your new threshold values for the -cpu and -mem options as marked in bold in the below example and press the return key:
perl /opt/UDoctor/FE/bin/sys_monitor.pl -mon -cpu 90 -mem 90
-
Select the SAVE|QUIT option by typing 7 into the interface and press the return key to save your changes and quit the UDoctor interface.
-
Upon exiting the interface, verify that the displayed configuration output matches your changes:
[sys_monitor] ENABLE = yes COMMAND = perl /opt/UDoctor/FE/bin/sys_monitor.pl -mon -cpu 45 -mem 45 SCHEDULE_TYPE = Minutes SCHEDULE_TIME = 15 GROUP = monitoring CLOUD_DEPENDENT = no
(The first one is for "sys_monitor," and the second one is for "sys_monitor_resolve" respectively.)
perl /opt/UDoctor/FE/bin/sys_monitor.pl -mon -cpu 90 -mem 90 perl /opt/UDoctor/FE/bin/sys_monitor.pl -res -cpu 85 -mem 85
To disable the monitoring feature
-
Log in to the array’s primary SP.
-
Disable the sys_monitor and sys_monitor_resolve jobs using the UDoctor service wrapper script by running the following commands, respectively:
svc_udoctor -j -d sys_monitor svc_udoctor -j -d sys_monitor_resolve
To add email messages to the user notification list:
See Dell KB article 22531: Dell EMC Unity: How to add SMTP server settings for Unity array. (User Correctable)
- A minimum severity level of WARNING is required for receiving the initial high usage alerts and a minimum severity level of NOTICE for receiving both high usage and resolve alerts.
- Popup notifications are still shown in the Unisphere UI when alerts are triggered, regardless if user email messages are added. The email notifications have additional information regarding the triggered alert such as CPU or Memory usage percentage values, as opposed to the UI pop-up notification.