[asterisk-dev] possible zaptel problem with SMP and RAID1

[François Delawarde]

No sharing here:

# cat /proc/interrupts
           CPU0       CPU1
  0:     247796  228080870   IO-APIC-edge      timer
  1:          0          8   IO-APIC-edge      i8042
  6:          0          3   IO-APIC-edge      floppy
  8:          0          0   IO-APIC-edge      rtc
  9:          0          0   IO-APIC-fasteoi   acpi
 12:          0          4   IO-APIC-edge      i8042
 14:          1         62   IO-APIC-edge      libata
 15:          0          0   IO-APIC-edge      libata
 17:  228158120     539814   IO-APIC-fasteoi   wctdm
 18:       5082   18200039   IO-APIC-fasteoi   eth2, eth0
 19:          0          0   IO-APIC-fasteoi   libata
 21:      38650   25145159   IO-APIC-fasteoi   eth1
 22:        166       5856   IO-APIC-fasteoi   libata
 23:       7811    2701437   IO-APIC-fasteoi   libata
NMI:          0          0
LOC:  228295439  228295357
ERR:          0

but counting interrupts give something like the following when no 
network traffic:

# INIT=`cat /proc/interrupts | grep wctdm | awk '{ print $2+$3; }'`; 
sleep 10; END=`cat /proc/interrupts | grep wctdm | awk '{ print $2+$3; 
}'`; echo $END-$INIT | bc
10007

and the following with 10Mbps, which by the way is not that much:

# INIT=`cat /proc/interrupts | grep wctdm | awk '{ print $2+$3; }'`; 
sleep 10; END=`cat /proc/interrupts | grep wctdm | awk '{ print $2+$3; 
}'`; echo $END-$INIT | bc
11203


Could it be a zaptel failure (I don't see how the card itself could 
start generating more interrupts based on quantity of network traffic)? 
If so, could someone indicate me how timing work and where to look in 
the source to try and identify the problem?

Also could it be a card failure (knowing that I tried with both Openvox 
and Digium cards)? I heard from unreliable sources that the "tiger jet 
320" chip on TDM400P and similar card wasn't that reliable. Could it be 
a hardware issue with it, or with a cheap crystal on the cards i tested? 
In that case, is it worth trying Sangoma, as some are saying that it is 
more reliable at IRQ level and work with more motherboards?


Thanks again,
François.


Martin Vít wrote:
> François Delawarde wrote:
>   
>> Hello,
>>
>> I've done some further investigation about the zaptel interrupt issue on 
>> my systems and found out the following:
>>
>> - Disabling software RAID1 didn't help, so the whole issue I had brought 
>> wasn't one.
>>   
>>     
> so my words confirms this :-)
>
>   
>> - I found out that it was related to network traffic 
>>     
> My friend was implementing E1 server which have some onboard ethernet 
> controler and there was problem asap the traffic on this ethernet starts 
> because of some IRQ sharing which cannot be changed (NOAPIC etc. doesnt 
> helped).
>
> _______________________________________________
> --Bandwidth and Colocation Provided by http://www.api-digital.com--
>
> asterisk-dev mailing list
> To UNSUBSCRIBE or update options visit:
>    http://lists.digium.com/mailman/listinfo/asterisk-dev