Matthieu Bec wrote:
>
> The system is an mv167 / 712 transition module (10baseT) / vxworks 5.2b,
> an average size EPICS database (500 records) under 3.12.
>
> ei (unit number 0):
> Flags: (0x8063) UP BROADCAST MULTICAST ARP RUNNING
> Type: ETHERNET_CSMACD
> Internet address: 10.2.2.121
> Broadcast address: 10.2.2.255
> Netmask 0xff000000 Subnetmask 0xffffff00
> Ethernet address is 08:00:3e:25:53:72
> Metric is 0
> Maximum Transfer Unit size is 1500
> 0 octets received
> 0 octets sent
> 2534236 packets received
> 12661 packets sent
> 2534236 unicast packets received
> 4294967290 unicast packets sent
> 0 non-unicast packets received
> 12667 non-unicast packets sent
> 0 input discards
> 0 input unknown protocols
> 24186 input errors
> 6 output errors
> 0 collisions; 0 dropped
>
> First I was reasonably impressed with `4294967290 unicast packets sent`
> (system had been up for 20h), but looking closely, I read this as a
> negative number of packets. Not being too familiar with the concept, I
> thought only the particle physics community would come up with an
> answer.
> 4294967290 apparently being (unsigned)-6, did I roll over my counters,
> or is it just a secondary artifact and I should focus on the errors
> reported in the first place?
The other related item is that you have 6 output errors. Tornado seems to
adjust the wrong packet counter when there is an output error, either that
or it increments the wrong packet counter when it sends one out, but in
either case the two cancel out. Apparently there are also multiple
counters, and doing an ifShow() from the host-based shell gives different
results to the target shell.
> Any hindsight on what might be causing this (or any good network
> reference to consult) would be greatly appreciated. As expected, the
> 'singularity' punctually happened, and went away.
Ignore the "funny" and keep an eye your input error counts, which are
probably real faults. If that number keeps on incrementing check whether
you're seeing similar errors at the other end.
- Andrew
--
Larry McVoy: "Learn how to think in C++ but don't ever program in it."
- References:
- making sense of peculiar ifShow report Matthieu Bec
- Navigate by Date:
- Prev:
making sense of peculiar ifShow report Matthieu Bec
- Next:
Missing SizeOffset Routine Dennis Armstrong
- Index:
1994
1995
1996
1997
1998
1999
2000
2001
<2002>
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
- Navigate by Thread:
- Prev:
making sense of peculiar ifShow report Matthieu Bec
- Next:
Missing SizeOffset Routine Dennis Armstrong
- Index:
1994
1995
1996
1997
1998
1999
2000
2001
<2002>
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
|