Highlighted
jcandiello
2 Iron

How can I check VNXe processor performance?

Within Unisphere, go to System > Processor Performance.

The CPU Usage History chart displays I/O performance statistics for Storage Processors (SPs) in your system. The statistics are color-coded for each SP. Beneath the graph, click the check box next to each SP to hide or show the line graph for a particular SP. Mouse-over a line graph to view the usage percentage and date when the performance statistics were collected.

Chart Components and their Descriptions

Percentage: Percentage of SP utilization up to 100%

Duration: Date and time over the specified interval.

Timeframe: Click to select a time period: hourly, last 24 hours, or last 48 hours.

1 Reply
asafayan1
1 Nickel

Re: How can I check VNXe processor performance?

Hi - you can also accomplish this via the CLI as follows:

CPU utilization graphs need to be generated for each service processor.

A. Establish ssh to VNXe - you will access either spa or spb
B. cd to cd /var/log/sa
Example:
service@(none) spb:/var/log/sa> ls -l
total 27716
-rw-r--r-- 1 root root 1388640 Oct  1 23:50 sa01
-rw-r--r-- 1 root root 1388640 Oct  2 23:50 sa02
-rw-r--r-- 1 root root  974120 Oct  3 16:40 sa03
-rw-r--r-- 1 root root 1388640 Sep 25 23:50 sa25
-rw-r--r-- 1 root root 1388640 Sep 26 23:50 sa26
-rw-r--r-- 1 root root 1388640 Sep 27 23:50 sa27
-rw-r--r-- 1 root root 1388640 Sep 28 23:50 sa28
-rw-r--r-- 1 root root 1388640 Sep 29 23:50 sa29
-rw-r--r-- 1 root root 1388640 Sep 30 23:50 sa30
-rw-r--r-- 1 root root 1902887 Oct  1 23:59 sar01
-rw-r--r-- 1 root root 1902887 Oct  2 23:59 sar02
-rw-r--r-- 1 root root  952154 Oct  3 11:59 sar03
-rw-r--r-- 1 root root 1902887 Sep 25 23:59 sar25
-rw-r--r-- 1 root root 1902887 Sep 26 23:59 sar26
-rw-r--r-- 1 root root 1902887 Sep 27 23:59 sar27
-rw-r--r-- 1 root root 1902887 Sep 28 23:59 sar28
-rw-r--r-- 1 root root 1902887 Sep 29 23:59 sar29
-rw-r--r-- 1 root root 1902887 Sep 30 23:59 sar30

C. less sarXX where XX is two digit number correlating to day of the month
Example:
This shows values for spb:
service@(none) spb:/var/log/sa> less sar01
Linux 2.6.32.59-0.7.1.00.036.NeoMain.SLES11SP1.NEO (spb)        2014-10-01      _x86_64_

00:00:01        CPU      %usr     %nice      %sys   %iowait    %steal      %irq     %soft    %guest     %idle
00:10:01        all      9.53      0.01      4.51      0.51      0.00      0.01      0.30      0.00     85.12
00:10:01          0     12.72      0.02      8.64      0.49      0.00      0.00      0.04      0.00     78.10
00:10:01          1     14.48      0.02      4.63      0.87      0.00      0.00      0.16      0.00     79.84
00:10:01          2      6.09      0.01      2.90      0.48      0.00      0.05      0.23      0.00     90.25
00:10:01          3      4.85      0.00      1.88      0.22      0.00      0.00      0.76      0.00     92.29

D. Issue "ssh peer" to access other Service Processor
Example:
service@(none) spb:/var/log/sa> ssh peer
Last login: Wed Oct  1 20:05:25 2014 from peer
service@(none) spa:~> cd /var/log/sa
service@(none) spa:/var/log/sa> ls -l
total 26452
-rw-r--r-- 1 root root 1296480 Oct  1 23:50 sa01
-rw-r--r-- 1 root root 1296480 Oct  2 23:50 sa02
-rw-r--r-- 1 root root  918480 Oct  3 16:50 sa03
-rw-r--r-- 1 root root 1296480 Sep 25 23:50 sa25
-rw-r--r-- 1 root root 1296480 Sep 26 23:50 sa26
-rw-r--r-- 1 root root 1296480 Sep 27 23:50 sa27
-rw-r--r-- 1 root root 1296480 Sep 28 23:50 sa28
-rw-r--r-- 1 root root 1296480 Sep 29 23:50 sa29
-rw-r--r-- 1 root root 1296480 Sep 30 23:50 sa30
-rw-r--r-- 1 root root 1844389 Oct  1 23:59 sar01
-rw-r--r-- 1 root root 1844383 Oct  2 23:59 sar02
-rw-r--r-- 1 root root  922962 Oct  3 11:59 sar03
-rw-r--r-- 1 root root 1844321 Sep 25 23:59 sar25
-rw-r--r-- 1 root root 1844377 Sep 26 23:59 sar26
-rw-r--r-- 1 root root 1844301 Sep 27 23:59 sar27
-rw-r--r-- 1 root root 1844357 Sep 28 23:59 sar28
-rw-r--r-- 1 root root 1844321 Sep 29 23:59 sar29
-rw-r--r-- 1 root root 1844437 Sep 30 23:59 sar30
service@(none) spa:/var/log/sa>

E. Repeat above steps to gather CPU metrics for service processor A.

NOTE:
Time interval is very 10 minutes for each CPU CORE 0,1,2 and 3.

NOTE:
Deduct %idle from 100 to derive CPU utilization

0 Kudos