[nos-bbs] sync timer testing

[Michael Fox - N6MEF]

Maiko,

 

I set the sync timer to 250 ms on all three ports.  Trace level was set to
0111 on all three ports.

 

I sent back-to-back 10KB messages into the BBS.  These were messages
destined for another BBS.  So, as soon as the first one was received, it was
forwarded to the next BBS on a different port while the next message was
being received, and so on.  In other words, two ports were running pretty
much continuously.

 

While that was happening, I monitored the CPU and memory activity using
Ubuntu System Monitor and saw no appreciable increase over the no traffic
state.  I'm not sure what other tools would be good for monitoring stuff
like this. 

 

Next, I tried 100 ms sync times on all three ports and again sent through
the same type of traffic.  Again, no appreciable change in the cpu, memory
or swap stats.

 

While experimenting, I notice that shorter intervals (< 1 sec) are useful
for interactive monitoring.  For example, it is nice to see the log entry at
nearly the same time that the radio s-meter is moving in order to know which
user is registering on the meter.   For longer term measurements, I would
think that 1 sec is a sufficient interval and I think I'll suggest that we
leave it set at 1 sec.  But the system appears to be very robust down to at
least 0.1 seconds. 

 

Nice work!

 

Michael

N6MEF

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.tapr.org/pipermail/nos-bbs_lists.tapr.org/attachments/20101110/bbd3e833/attachment.html>