Wired Intelligent Edge

 View Only

  • 1.  ArubaCX. "Top cpu" over 100%

    Posted Feb 16, 2022 08:00 AM
      |   view attached
    Good day.
    I inherited network build around aruba 8320
    Version : TL.10.06.0001
    Build Date : 2020-11-10 10:31:42 PST
    Build ID : ArubaOS-CX:TL.10.06.0001:55dffa340d0f:202011101711
    Build SHA : 55dffa340d0fe49fb9088928e34b71a48e32f80e
    Active Image : primary
    Service OS Version : TL.01.05.0003
    BIOS Version : TL-01-0013

    Could you help me understand her cpu utilization? (screen in attachment)

    1)Is it a single core device and load average > 1 means it is saffocating?
    2)Why "top" itself consumes 37.5% cpu? 

    tnx for help.




    ------------------------------
    Alexander Progcom
    ------------------------------


  • 2.  RE: ArubaCX. "Top cpu" over 100%

    Posted Feb 16, 2022 08:31 AM
    Hi Alexander, to see all CPUs during the top session you need to press "1". This is what I see on Aruba 8320 (AOS-CX 10.7.

    switch:~$ top
    top - 14:25:34 up 61 days, 23:30, 1 user, load average: 0.83, 1.29, 1.23
    Tasks: 250 total, 2 running, 248 sleeping, 0 stopped, 0 zombie
    %Cpu0 :  9.4 us, 3.4 sy, 0.0 ni, 84.9 id, 0.0 wa, 0.0 hi, 2.3 si, 0.0 st
    %Cpu1 : 16.9 us, 3.7 sy, 0.0 ni, 77.4 id, 0.0 wa, 0.0 hi, 2.0 si, 0.0 st
    %Cpu2 :  4.7 us, 3.0 sy, 0.0 ni, 89.5 id, 0.0 wa, 0.0 hi, 2.7 si, 0.0 st
    %Cpu3 :  8.3 us, 3.6 sy, 0.0 ni, 85.4 id, 0.3 wa, 0.0 hi, 2.3 si, 0.0 st
    MiB Mem : 15843.7 total, 9402.2 free, 1690.3 used,  4751.2 buff/cache
    MiB Swap:     0.0 total,    0.0 free,    0.0 used. 12395.7 avail Mem

    PID  USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
    1938 root 20 0 875924 333120 96284 S 20.5 2.1 19733:08 ops-switchd
    864 root 20 0 35896 21112 5436 R 11.9 0.1 11342:11 ovsdb-server
    1231 root 20 0 695024 31836 13700 S 4.6 0.2 4390:26 ndmd
    1959 root 20 0 129044 14912 10028 D 3.0 0.1 2861:45 pmd
    1779 root 20 0 5480956 104116 39816 S 2.6 0.6 2601:49 prometheus
    1776 root 20 0 1254916 75560 16748 S 1.3 0.5 1763:04 hpe-tsdbd
    14607 root 20 0 356852 63796 29336 S 1.3 0.4 1157:48 hpe-routing
    1168 root 20 0 223384 66324 10172 S 1.0 0.4 526:37.83 hpe-pvstd
    10 root 20 0 0 0 0 I 0.7 0.0 562:40.35 rcu_preempt
    1216 root 20 0 92088 17496 11120 S 0.7 0.1 159:20.83 hpe-mstpd
    1266 root 20 0 85288 12524 9108 S 0.7 0.1 565:28.29 hpe-vrrpd
    CPU usage varies.

    I suggest you to plan an upgrade from AOS-CX 10.06 to 10.07 or 10.08.

    ------------------------------
    Davide Poletto
    ------------------------------

    Original Message


  • 3.  RE: ArubaCX. "Top cpu" over 100%

    Posted Feb 16, 2022 08:41 AM
    Edited by alext Feb 16, 2022 08:50 AM
    Unfortunately pressing 1 acts like any other keys and just displays more rows
    Standalone top is not supported.
    8320# top
    % Command incomplete.

    I`ll plan OS upgrade. Tnx

    And could you give a hint why 
    Why "top" itself consumes 37.5% cpu? 


  • 4.  RE: ArubaCX. "Top cpu" over 100%

    Posted Feb 16, 2022 09:43 AM
    Edited by Ivan_B Feb 16, 2022 09:53 AM
    Hi Alexander,

    'top cpu' command provides you pretty much similar output as the 'top' command you can find on Linux, however it lacks many interactive features of the  'real' top command, that is why pressing '1' doesn't give you expected result. 

    Use 'start-shell' command in order to access Linux Bash shell and from there strart the full-blown 'top'. There you will have all the features described in the manual  - https://man7.org/linux/man-pages/man1/top.1.html

    (In fact the 'top cpu' in AOS-CX shell is like 'top -b -n 1' in Linux)

    To the question why the 'top cpu' command consumes 37% CPU. The answer is pretty simple - that 37% was utilization of one core at the moment when the command is executed. After all the 'top' should start itself, poll a lot of /proc variables etc etc... If you enter the Bash shell and start the interactive 'top' command you will see the similar picture during the first iteration, but after 3 seconds when the process list refreshes again, the utilization of 'top' process will go down significantly. 


    ------------------------------
    Ivan Bondar
    ------------------------------

    Original Message


  • 5.  RE: ArubaCX. "Top cpu" over 100%

    Posted Feb 16, 2022 10:05 AM
    Edited by Ivan_B Feb 16, 2022 10:17 AM
    Alexander,

    If we need to analyze the state of your 8320 CPU, I'd monitor the 'load average' output of the 'top cpu' (or 'top' in Bash shell) command. That tells much more about the overall state of your CPU than momentary snapshot that the process list gives you. In your case everything is good. You have 4 cores, thus any number below 4.0 means your CPU is under-utilized. 4.0 will mean 400% which on its turn means all 4 cores at 100%. Keep in mind that unlike 'cpu utilization' parameter that can't go higher than 400% in your case, the 'load average' can be higher. And if it will become higher, it will mean that you have issue with CPU over-utilization and some processes start to wait for the CPU to execute them. More on the 'load average' indicator you can read here - https://man7.org/linux/man-pages/man1/uptime.1.html

    In your case the 'load average' numbers are "1.26, 1.03, 1.00". From left to right, these numbers show you the average load over the last one minute, the last five minutes, and the last fifteen minutes.  Which translates to:

    Over the last 1 minute 4 CPU cores were loaded about 32% of the time
    Over the last 5 minutes 4 CPU cores were loaded about 26% of the time
    Over the last 15 minutes 4 CPU cores were loaded 25% of the time


    ------------------------------
    Ivan Bondar
    ------------------------------

    Original Message


  • 6.  RE: ArubaCX. "Top cpu" over 100%

    Posted Feb 16, 2022 03:43 PM
    Tnx a lot.
    start-shell allowed to view normal "top" and no more confusion.

    ------------------------------
    Alexander Progcom
    ------------------------------

    Original Message


Related Content