CPU - Interrupts distribution

Applications Infrastructure Monitoring
1

Hi Gurus,

I have a situation runing my ETL tools on the below server. The throughput while processing data is coming very low. When i tried to analyse the CPU stats i got colleceted the

mpstat

.

Server Physical Host Name *********com

IP**.***.**
OS Type Linux
OS Classlinux red hat enterprise linux
OS Versionlinux system x3550 m2 -[7946ac1]-
MS Build NameMSSB-LEGACY
MS Build VersionKernel2.6.18-247.el5.bz683155
Kernel Architecturex86_64
Manufacturerlinux
Modelsystem x3550 m2 -[7946ac1]-Server Config2-WayServer ClassNon-Blade
Model RollupSystem x3550 M2 -[7946AC1]-
Number Of CPU2
Number Of CPU Cores2
Virtual CPU Count16CPU Type2.22 GHz x86_64CPU 
Speed2220RAM MB48271
Tier DescriptionTier 3 - Restricted Availability
Internal Disk SizeNASAN IdNA

i have attached the

mpstat

info i collected.
My worry is whenever the interrupts / sec > 3000 , the interrupts are distributed to cpu0 and cpu 15 for major chunk , rest of the cpu's are sitting idle. Even in that distribution CPU 0 only gets 1000 interrupts and CPU 15 has the remaining ~ 80 -90% of the load.
Is the behaviour correct , kindly provide an opinion,

2

PFA the

mpstat

output

3

Also captured the yeild for

cat /proc/interrupts

           CPU0,    CPU1,     CPU2,     CPU3,   CPU4,    CPU5,  CPU6,    CPU7,    CPU8,    CPU9,    CPU10,      CPU11,  CPU12,    CPU13,    CPU14,   CPU15,
130:00:00,  1048,       0,      0,       0,       0,      0,       0,       0,       0,       0,        0,  202064668,       0,        0,        0, 512532716, PCI-MSI, eth0 
185:00:00, 5385, 3181749, 3266781, 1267992, 3453589, 2981991, 104657, 265, 0, 66125, 10136, 85428, 2100126, 1341219, 65914, 0, IO-APIC-level, lpfc
193:00:00, 4453, 3421072, 3176832, 975485, 2866437, 3487992, 302016, 29507, 0, 960, 67873, 100243, 1338298, 2182582, 0, 0, IO-APIC-level, lpfc
201:00:00, 7326, 21563193, 12967850, 1946063, 0, 62303, 5370651, 543112, 0, 3329977, 14479511, 65730, 0, 19, 12861008, 20370, IO-APIC-level, megasas
NMI:, 323299, 99906, 73879, 61131, 266359, 91899, 75309, 61200, 102974, 104665, 80748, 104176, 98485, 119343, 91712, 114415, , 
LOC:, 4155248420, 4155248563, 4155248841, 4155248725, 4155248852, 4155248711, 4155249174, 4155249117, 4155248301, 4155247903, 4155248172, 4155248183, 4155248709, 4155248296, 4155248535, 4155248532,

saw uneven
distribution for CPU 0 , CPU 11 and CPU 15 , can anyone suggest if this normal behaviour ?

4
Moderator comments were removed during original forum migration.
1 Like