Re: 20011031: LDM latency

Patrick O'Reilly wrote:
> 
> Anne,
> 
> I have used notifyme on papagayo, and it always seems to be pretty much on
> time.  I double check the IDD latency stats on the web and that verifies
> what I see.  I ran 'netcheck' and it basically is a script that runs 'ping'
> and 'traceroute', right?  I have run those seperately, but with my limited
> knowledge of what the output tells me, I can't make too much of them.  Below
> is the output of 'netcheck from blizzard, some packet loss (10%) showing
> from the 'ping' program:
> 
> [ldm@blizzard]#./netcheck
> 
> netcheck                       Nov 5 22:31:36 UTC                Version:
> 1.02
> ============================================================================
> ===
> 
> Upstream Hosts:
> 
> Nov 5 22:32:45 UTC: papagayo.unl.edu
> ________________________
> 
> ping -s papagayo.unl.edu 64 10:
> ----papagayo.unl.edu PING Statistics----
> 10 packets transmitted, 9 packets received, 10% packet loss
> round-trip (ms)  min/avg/max = 93/107/126
> 
> traceroute papagayo.unl.edu:
> traceroute to papagayo.unl.edu (129.93.52.150), 30 hops max, 40 byte packets
>  1  starbase1.uninet.uni.edu (134.161.128.1)  0.490 ms  0.375 ms  0.344 ms
>  2  uni-gate.uni.edu (134.161.1.101)  6.635 ms  3.587 ms  6.315 ms
>  3  * * *
>  4  * * *
>  5  * sl-gw12-che-2-0.sprintlink.net (160.81.219.137)  28.155 ms  43.584 ms
>  6  sl-bb21-che-3-2.sprintlink.net (144.232.15.173)  103.210 ms  209.354 ms
> 36.
> 413 ms
>  7  sl-bb21-chi-11-2.sprintlink.net (144.232.18.6)  46.953 ms  45.280 ms
> 158.21
> 8 ms
>  8  sl-gw11-chi-0-0-0.sprintlink.net (144.232.10.110)  101.133 ms  51.904 ms
> 60
> .833 ms
>  9  sl-lincoln-2-0-0-T3.sprintlink.net (144.228.156.34)  129.709 ms  183.674
> ms
>  149.325 ms
> 10  162.40.115.10 (162.40.115.10)  100.042 ms  157.630 ms  120.488 ms
> 11  205.240.124.78 (205.240.124.78)  102.999 ms  106.887 ms *
> 12  * * *
> 13  papagayo.unl.edu (129.93.52.150)  90.097 ms  103.028 ms  99.864 ms
> 
> Downstream Hosts:
> 
> [ldm@blizzard]#
> 
> Hope this gives you a few clues.  I have written my sys admin guy about
> network issues, and have received no reply.  They are obviously not nearly
> as supportive as Unidata, as I can see!  When you have time, let me know

Hi Patrick,

Maybe the end of your message got cut off there...

Hops 8 - 11 in your traceroute that don't look that good.  

But, what I was proposing, which I didn't make clear, is that you run
netcheck continually for a day or two.  That is, make a cron job for
it.  Here's an example from my ldm's crontab (it is currently commented
out):

# run netcheck every 2 minutes
#*/2 * * * *    /usr/local/ldm/bin/netcheck


Have netcheck log to a file, like ~ldm/logs/netcheck.log.  Then we can
get a view of the connectivity over time.

Anne
-- 
***************************************************
Anne Wilson                     UCAR Unidata Program            
address@hidden                 P.O. Box 3000
                                  Boulder, CO  80307
----------------------------------------------------
Unidata WWW server       http://www.unidata.ucar.edu/
****************************************************