Re: Sudden Backup Slowdown

Hi Nenad!

The problem with this little program is that it will execute completely from within the CPU cache, which will make the result unrealistically fast.

Regards

On 11/5/19 7:16 AM, Noveljic Nenad wrote:
>
> Hi Charlotte,
>
> The little C program I sent you previously is for Solaris – sorry. The
> following version compiles on Linux as well:
>
> #include<time.h>
>
> #include <sys/time.h>
>
> #include <stdio.h>
>
> long get_hr_timestamp_ns();
>
> int main()
>
> {
>
>   long start, end;
>
>   int i, j;
>
>   printf("loop duration [ns]:\n");
>
> printf("------------------------------------\n");
>
>   while(1){
>
>     start = get_hr_timestamp_ns();
>
>     for ( i = 0; i < 1e8; i++ )
>
>     {
>
>       j = j + 1;
>
>     }
>
>     end = get_hr_timestamp_ns();
>
>     printf("%ld\n", end - start );
>
>   }
>
> }
>
> long get_hr_timestamp_ns()
>
> {
>
>   long t ;
>
> #if defined(__sun)
>
>   t = (long)gethrtime();
>
> #endif
>
> #if defined(__linux__)
>
>   struct timespec ts;
>
>   clock_gettime(CLOCK_MONOTONIC, &ts);
>
> t = ts.tv_sec * 1e9 + ts.tv_nsec;
>
> #endif
>
>   return t;
>
> }
>
> This is how you compile it:
>
> gcc cpu_loop.c –o cpu_loop
>
> Best regards,
>
> Nenad
>
> https://nenadnoveljic.com/blog/
>
> *From:*Noveljic Nenad
> *Sent:* Dienstag, 5. November 2019 10:28
> *To:* 'Charlotte Hammond' <charlottejanehammond_at_yahoo.com>;
> dmarc-noreply_at_freelists.org; ORACLE-L <oracle-l_at_freelists.org>
> *Subject:* RE: Sudden Backup Slowdown
>
> Hi Charlotte,
>
> So, on the OS you're seeing effectively 32 CPUs?
>
> By the way, there must be 4 to 9 other processes competing with
> backups for the CPUs. Maybe, that's something worth looking into,
> because you didn't expect that (you mentioned that there's nothing
> else running concurrently), even though that's not the cause for
> increase in elapsed time (the run queue length over time is the same
> in both cases).
>
> Since the AWR statistics (blocks read, written, etc.) are the same for
> both cases we can also exclude the increased data volume as the cause
> of the increased elapsed time.
>
> I would, therefore, focus on the following:
>
> 1. Test CPU performance (as suggested by Dimitre):
>
> - single thread over time
>
> - increasing the load from 1 to #CPUs parallel threads to see if the
> performance of a single thread is decreasing as the load is increasing
> (it shouldn't).
>
> If you don't have a baseline from when it was running "fast" you could
> compare the performance to other VMs, preferably in different
> geographic locations, as suggested in the article posted by Dimitre.
>
> You can use either some standard performance tool or simply measure
> the execution time of a simple C loop, like this:
>
> #include<time.h>
>
> #include <sys/time.h>
>
> #include <stdio.h>
>
> int main()
>
> {
>
>   long start, end;
>
>   int i, j;
>
>   while(1)
>
>   {
>
>     start = (long)gethrtime();
>
>     for ( i = 0; i < 1e8; i++ )
>
>     {
>
>       j=i+1 ;
>
>     }
>
>     end = (long)gethrtime();
>
>     printf("ns: %ld\n", end - start );
>
>   }
>
> }
>
> I'm aware that your SLAs are mentioning a guaranteed performance.
> Nevertheless, I think it's worth double-checking in such a situation.
>
> 2. Profiling stacks
>
> If the CPU performance hasn't changed, then you could sample stacks
> and profile them. Even though we can't compare it to before, we can
> see if some unexpected functions appear on the stack.
>
> Best regards,
>
> Nenad
>
> https://nenadnoveljic.com/blog/
>
> *From:*Charlotte Hammond <charlottejanehammond_at_yahoo.com
> <mailto:charlottejanehammond_at_yahoo.com>>
> *Sent:* Dienstag, 5. November 2019 09:28
> *To:* dmarc-noreply_at_freelists.org
> <mailto:dmarc-noreply_at_freelists.org>; ORACLE-L <oracle-l_at_freelists.org
> <mailto:oracle-l_at_freelists.org>>; Noveljic Nenad
> <nenad.noveljic_at_vontobel.com <mailto:nenad.noveljic_at_vontobel.com>>
> *Subject:* Re: Sudden Backup Slowdown
>
> Hi Nenad,
>
> The average run queue is about 20 (peaking at about 25) whilst the
> backup is running.   The figure is the same for both the slow and fast
> runs.   I've graphed them out and they look pretty much the same.  
> This is a fairly high figure but it doesn't seem to have changed.
>
> We have 16 backup channels and 16 CPUs (2 threads per processor).
>
> Yes, the compression is "basic" but that has not changed.
>
> Thanks,
>
> Charlotte
>
> ____________________________________________________
>
> Please consider the environment before printing this e-mail.
>
> Bitte denken Sie an die Umwelt, bevor Sie dieses E-Mail drucken.
>
>
> Important Notice
>
> This message is intended only for the individual named. It may contain
> confidential or privileged information. If you are not the named
> addressee you should in particular not disseminate, distribute, modify
> or copy this e-mail. Please notify the sender immediately by e-mail,
> if you have received this message by mistake and delete it from your
> system.
> Without prejudice to any contractual agreements between you and us
> which shall prevail in any case, we take it as your authorization to
> correspond with you by e-mail if you send us messages by e-mail.
> However, we reserve the right not to execute orders and instructions
> transmitted by e-mail at any time and without further explanation.
> E-mail transmission may not be secure or error-free as information
> could be intercepted, corrupted, lost, destroyed, arrive late or
> incomplete. Also processing of incoming e-mails cannot be guaranteed.
> All liability of Vontobel Holding Ltd. and any of its affiliates
> (hereinafter collectively referred to as "Vontobel Group") for any
> damages resulting from e-mail use is excluded. You are advised that
> urgent and time sensitive messages should not be sent by e-mail and if
> verification is required please request a printed version.
> Please note that all e-mail communications to and from the Vontobel
> Group are subject to electronic storage and review by Vontobel Group.
> Unless stated to the contrary and without prejudice to any contractual
> agreements between you and Vontobel Group which shall prevail in any
> case, e-mail-communication is for informational purposes only and is
> not intended as an offer or solicitation for the purchase or sale of
> any financial instrument or as an official confirmation of any
> transaction.
> The legal basis for the processing of your personal data is the
> legitimate interest to develop a commercial relationship with you, as
> well as your consent to forward you commercial communications. You can
> exercise, at any time and under the terms established under current
> regulation, your rights. If you prefer not to receive any further
> communications, please contact your client relationship manager if you
> are a client of Vontobel Group or notify the sender. Please note for
> an exact reference to the affected group entity the corporate e-mail
> signature. For further information about data privacy at Vontobel
> Group please consult www.vontobel.com <https://www.vontobel.com>.
>

-- 
Mladen Gogala
Database Consultant
Tel: (347) 321-1217


--
http://www.freelists.org/webpage/oracle-l