3par allocated space thresholds

[Tienna]

Hi Gurus

I am trying to monitor 3par with 3par alerts, I want to set system send an email whenever allocated space is over 90%. but this threshold is not available in the list of SSMC. Can any one guide me to create once.

Thank you

[Sheldon Smith]

SSMC thresholds deal with performance thresholds.
You want to set a Warning value on the desired object (virtual volume or CPG). The Service Processor is in charge of things like capacity alerts..

Make sure the Service Processor is configured to send email alerts. Then the following will generate a test capacity alert using a test CPG.

---

Summary: Customers wishing to test their notification (SNMP and/or local email) setup can create a critical alert via a createcpg CLI command. The CPG is created with an allocation limit of 1G. Since the minimum growth of a CPG is 8G, the CPG grows to 1G in size and immediately reaches the allocation limit. Following the alert testing, the CPG, corresponding LDs and alert can be removed.

First check the system for any outstanding alerts:

3PAR cli% showalert

See if the list is either empty, or any alerts have been "resolved by system".

Then: Create the CPG with a 1GB growth limit:

createcpg -sdgl 1g test_alert_cpg

However for this CPG to grow, the customer would also have to create a VV with the same size as the CPG. The CPG will then reach its allocation limit. We just need to create a minimal VV:

createvv test_alert_cpg test_alert_vv 1g

An alert will then be triggered:

3PAR cli% showalert

State       : New

Message Code: 0x270006

Severity    : Critical

Type        : CPG growth limit

Message     : CPG test_alert_cpg SD space has reached allocation limit of 0.0009766G

After the testing has been verified the customer can remove the VV by issuing the following command:

removevv -f test_alert_vv

and remove the remove the CPG by issuing:

removecpg -f test_alert_cpg 

[Dennis Handly]

There are also raw space alerts based in device type that can be set with setsys.

You can see them with "showsys -param".

[Sheldon Smith]

True enough, and since the system's primary resource is free Raw Space, that is, free chunklets, keeping an eye on the free space has always made more sense to me than seeing how full a particular volume may be. Which seems more a task for the host's sysadmin. I guess it depends on how many hats a person has to wear. ;)

Note that there are default raw space alerts for each device-type (NL, FC, SSD) at

• 50% - informational
• 75% - warning
• 85% - severe
• 95% - critical