Hi Berthold,

Actually, it's not hard at all. Well, provided powertop would work, especially 
the v1.13 version. Powertop shows the wakeup of programs/processes. I have used 
this in all Linux guests in the past to tune the systems. I also have used this 
in my openSUSE PC, though obviously X-windows and USB already give a lot of 
timer wakeups in there. I can't run v1.13 anymore in any Suse system, it looks 
like a specific option is no longer compiled into the kernel. In my openSUSE 
system powertop v2 doesn't show the processes too clear, in my guest it doesn't 
show the processes at all.

For instance, on my TSM server (SLES11) powertop 1.13 shows:
Wakeups-from-idle per second : 46.8     interval: 5.0s

Top causes for wakeups:
  35.9% ( 32.6)   db2sysc
  24.0% ( 21.8)   [kernel scheduler] Load balancing tick
  17.2% ( 15.6)   [kernel core] hrtimer_start (tick_sched_timer)
  13.0% ( 11.8)   [kernel core] run_timer_softirq (garp_join_timer)
   6.2% (  5.6)   kworker/0:0
   1.1% (  1.0)   lin_taped
   1.1% (  1.0)   db2fmcd

So here we have 46.8 wakeups per second, about 36% is caused by a db2 process. 
So every 6 miliseconds db2 wakes up leaving the guest in Q3 all the time. As 
for CPU, db2sync is never listed in top so you wouldn't see that in any other 
way. This at least show what service needs to be investigated. (Well, I wasn't 
able to lower the db2 wakeups enough to get the guest drop from queue.)

I have tried readprofile but it doesn't show me what I need. Indeed you can see 
a lot of kernel calls but that doesn't point to a specific program or service 
that I could investigate further. Calls to system or kernel processes are not 
helpful as any program can call them. If anything, enabled_wait is the call 
that is executed the most but it's not clear who is responsible for this. Also 
the numbers are meaningless when it's not related to elapsed time. You would 
want to see wakeup or interrupts per second.

Met vriendelijke groet/With kind regards/Mit freundlichen Grüßen,
Berry van Sleeuwen
Flight Forum 3000 5657 EW Eindhoven

-----Original Message-----
From: Linux on 390 Port <[email protected]> On Behalf Of Berthold Gunreben
Sent: Monday, 6 April 2020 17:51
To: [email protected]
Subject: Re: Performance monitoring SLES15

Hi Berry,

monitoring this kind of activity is typically quite hard. Often enough, the 
system might lie to you in terms of real activity. Thus here just one short 
recommendation regarding profiling kernel functions:

This e-mail and the documents attached are confidential and intended solely for 
the addressee; it may also be privileged. If you receive this e-mail in error, 
please notify the sender immediately and destroy it. As its integrity cannot be 
secured on the Internet, Atos’ liability cannot be triggered for the message 
content. Although the sender endeavours to maintain a computer virus-free 
network, the sender does not warrant that this transmission is virus-free and 
will not be liable for any damages resulting from any virus transmitted. On all 
offers and agreements under which Atos Nederland B.V. supplies goods and/or 
services of whatever nature, the Terms of Delivery from Atos Nederland B.V. 
exclusively apply. The Terms of Delivery shall be promptly submitted to you on 
your request.

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO LINUX-390 or visit
http://www2.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to