I've been trying to track down this type of loss for
quite some time..
what i've seen is
Solaris does not have this issue.
the free *nix'es do have this issue and it is worse
with more hosts being monitored.
there is the in_cksum() function in icmp.c. I should check
that to insure it is properly forming the checksum.
On Mon, Jun 10, 2002 at 11:50:42AM +0200, Bengt Gördén wrote:
> On Mon, Jun 10, 2002 at 09:37:24AM +0000, Simon White wrote:
> > 10-Jun-02 at 11:32, Bengt Gördén (email@example.com) wrote :
> > >
> > > We had sysmon running fine on FreeBSD and about 80 nodes to monitor
> > > with ping. We then moved sysmon to a NetBSD and put in 173 nodes to
> > > monitor with ping. The thing is that sometimes sysmon sends mail
> > > without there being an outage. I have monitored the log-files and even
> > > sent some pings myself when sysmon says there is an outage. My own
> > > pings went fine but sysmon reported outage. Anyone that have run into
> > > this problem or knows about related problems with NetBSD? Or have I
> > > missed something in the docs? I cant find any answer in the
> > > mail archive or in the docs.
> > I have had similar problems intermittently with RH Linux 7.1 (vanilla
> > install) but it appears to be OK now. The only reason I say this is that
> > it could be not NetBSD specific.
> Thanks for the info. Did you see any icmp bad checksums at the time
> when you had problems?
> Bengt Görden
-- Jared Mauch | pgp key available via finger from firstname.lastname@example.org clue++; | http://puck.nether.net/~jared/ My statements are only mine.
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:14:07 EDT