The problem is solved now. It turned out *not* to be the fault of
pfsense. Upgrading the firmware of my switches solved the problem -
mostly. Now the throughput fluktuates (instead of being steady at a high
rate): Peaks go up to 45MB/s, average is around 25-30MB/s (measured over
ca. 5 minutes) through the firewall (compared to 117MB/s with a direct
link). I observe some switching over to the slave CARP device when peak
speed is reached, which is the cause for the much lower average.
The settings of Base and Skew values ("Virtual IP Address: Edit") of the
slave device seem to be very delicate. Setting the base value too high
kills practically the CARP switching, a good value for the skew I don't
know. It seems to me that switching works good and fast with base=3
(around 3 seconds delay, intentionally?). For skew I set it to 50. What
are your experiences with this? If you flood your pfsense with traffic,
does it stay on the master or do you also see a switching to the slave?
What are your skew settings? Is your throughput at a steady rate?
Best regards, Adrian.
PS: Test has been done like the following, see original posting for details:
WAN machine: nc -l -p 10000 > /dev/null (Linux)
nc -l 10000 > /dev/null (BSD or OSX)
Local machine: cat /dev/zero | nc wan_machine 10000
On 7/16/13 6:26 PM, Adrian Zaugg wrote:
> Dear List
>
> This thread is still unsolved, any help would be very much appreciated.
>
> On 7/8/13 11:45 AM, Adrian Zaugg wrote:
>> What means "re3: watchdog timeout"? Is it possible that the watchdog
>> resets the interface?
>
> Please see the full message in the attachment.
>
> Regards, Adrian.
>
>
>
> _______________________________________________
> List mailing list
> List@lists.pfsense.org
> http://lists.pfsense.org/mailman/listinfo/list
>
_______________________________________________
List mailing list
List@lists.pfsense.org
http://lists.pfsense.org/mailman/listinfo/list
