I'm trying to determine how I can set up an alert - like if LUN 800 goes over 80% utilization or if LUN 1300 queue depth goes over 5, an email is sent.
Is there a method by which this can be done?
willing to do some scripting ?
naviseccli -Address vnx5700spa -Scope 0 -User sysadmin -Password blah lun -list -l 437
LOGICAL UNIT NUMBER 437
Name: LUN 437
Current Owner: SP B
Default Owner: SP B
Allocation Owner: SP B
User Capacity (Blocks): 20971520
User Capacity (GBs): 10.000 <-----------
Consumed Capacity (Blocks): 6303744
Consumed Capacity (GBs): 3.006 <----------
Pool Name: Pool 0
Raid Type: Mixed
Auto-Assign Enabled: DISABLED
Auto-Trespass Enabled: DISABLED
Current State: Ready
Is Faulted: false
Is Transitioning: false
Current Operation: None
Current Operation State: N/A
Current Operation Status: N/A
Current Operation Percent Completed: 0
Is Pool LUN: Yes
Is Thin LUN: Yes
Is Private: No
Is Compressed: No
Tiering Policy: Auto Tier
Initial Tier: Highest Available
Extreme Performance: 100.00%
I am not sure if you are running "VNX Monitoring and Reporting" i think you can setup some alerts in there but that will cover VNX, it does not support older platforms.
As far as getting queue length, i am not sure about that metric ..poke around in Navicli admin guide.
I guess I should have been more specific - when I say util%, I mean the Characteristic Utilization (%) as can be seen in the Performance Detail chart from NaviAnalyzer. Same for queue depth.
Not afraid of scripting, however, I would have thought there would be something in Unisphere/Navi UI.
I'll look through the admin guide.
Is there noone monitoring the performance of their LUNs at the SAN?
If there are, what do you use?
for my Clariion/VNX i am using Uni/Navi Analyzer , i archive them every week to an external directory and the delete old archives from SP. This way i can reference older NAR files if someones wants to go back and look at how application "A" performed 3 months ago. I think you are looking for something more pro-active, real-time ?
We just recently purchased a VNX5600 that comes with a free license for "VNX Monitoring and Reporting". It can monitor VNX2 and VNX gen arrays. I am actually impressed, it has good performance statistics.
Thanks again, dynamox.
Yep, looking for "realtime". Analyzer/NARs have a "large grained" interval (5min). The chart(s) have 1min. I'd prefer 1sec...
I can keep the Performance Detail screen up showing Utilization and Queue Length to get the "per minute" view "realtime", however, I can't watch it 24x7 . I'd rather have an email/page sent if one of the critical LUNs breaches our internal SLA.
Will "VNX Monitoring and Reporting" do this?
Is there anything that will do this for the NS-480 and/or CX3-80?
i have not played with VNX Monitoring and Reporting a whole but EMC offers a free 90 day license, install it and give a shot. I can't think of 3rd party tools for CX3/4 platform ..heard some folks talk about SolarWinds but don't have personal experience with it.
Realtime stats...that's going to be a tough one. I can't remember what the lowest collection value was for Analyzer, obviously it will add more load to CPUs and your archive files will roll-over much faster.
For what it's worth, the lowest collection value for Analyzer is 5min (300sec). This is why I was looking at the Performance Detail chart - it charts every minute. We have very tight externally facing SLAs here (5sec total time in our environment), thus every ms counts. I've noticed that (for the SAN) anytime we have a critical LUN go over 80% or that LUNs queue depth goes over 5, it indicates we might start to violate those timing SLAs. I was/am trying to head that off by sending alert(s) when that/those happen (try to get the machine to tell you something's wrong before something actually is...).
I'll take a look at "VNX Monitoring and Reporting".
It appears that version 1.1 includes (among other things) alerts; even a pre-defined alert for LUN response time:
Even states it's supported for the CX4 - very close - I'll try to test it.