[CentOS] Balancing IRQs

Tue Jul 29 10:00:02 UTC 2014
Jatin Davey <jashokda at cisco.com>

On 7/29/2014 12:43 PM, IP wrote:
> W dniu 2014-07-29 08:16, Jatin Davey pisze:
>> Hi All
>>
>> I am using the following Linux version for my application:
>>
>> ****************************************
>> [root at localhost ~]# uname -a
>> Linux localhost.localdom 2.6.18-164.el5 #1 SMP Thu Sep 3 03:28:30 EDT
>> 2009 x86_64 x86_64 x86_64 GNU/Linux
>> ****************************************
>>
>> ****************************************
>> [root at localhost ~]# cat /etc/*release
>> CentOS release 5.4 (Final)
>> ****************************************
>>
>> When my application is running on the server i can see only one core
>> being used heavily while the other cores are not being used much. My
>> application has a major workload is collecting data from the eth0
>> interface and then doing a lot of processing with the data in the
>> database. So i can say that the Disk I/O and Network I/O are the major
>> contributors towards the workload.
>>
>> from the /proc/interrupts file i can see the following:
>>
>> ****************************************************
>> [root at localhost ~]# cat /proc/interrupts
>>              CPU0       CPU1       CPU2       CPU3
>>     0:   16084980          0          0          0    IO-APIC-edge timer
>>     1:        141          0          0          0    IO-APIC-edge i8042
>>     6:          5          0          0          0    IO-APIC-edge floppy
>>     7:          0          0          0          0    IO-APIC-edge parport0
>>     8:          0          0          0          0    IO-APIC-edge rtc
>>     9:          0          0          0          0   IO-APIC-level acpi
>>    12:        251          0          0          0    IO-APIC-edge i8042
>>    51:    2410224          0          0          0   IO-APIC-level ioc0
>>    59:   11358257          0          0          0   IO-APIC-level eth0
>> NMI:          0          0          0          0
>> LOC:   16086147   16090336   16090320   16089691
>> ERR:          0
>> MIS:          0
>> ****************************************************
>>
>>   From the above output i can see that all the interrupts are hitting the
>> cpu0 and thats why it is used more than any other core in the system.
>>
>> Can someone please help me in distributing the interrupts across all the
>> cores so that all are used which will eventually result in a better
>> performance of my system.
>>
>> Appreciate your response in this regard.
> Hi.
>
> Just use service irq balance: irqbalance from base repo of centos 5
>
> IP
>   
>> Thanks
>> Jatin
>> _______________________________________________
>> CentOS mailing list
>> CentOS at centos.org
>> http://lists.centos.org/mailman/listinfo/centos
>>
> _______________________________________________
> CentOS mailing list
> CentOS at centos.org
> http://lists.centos.org/mailman/listinfo/centos
>
Thank you very much for the response.

Thanks
Jatin