[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

RE: WinXP and Valhalla machines do not respond to each other's ping's

> I don't trust <tongue-in-cheek> M$ networking </tongue-in-cheek> for
> anything outside the localhost, let alone allow it to have a
> routing table
> configured. Right now Phil doesn't have a dual-homed host set
> up, so all
> the eth interfaces are public. It would be a fairly complex
> set of routing
> tables.

And impossible to maintain reliably as his IP addresses are subject to
change with DHCP
upon the Whim of his ISP. he might get lucky and always release the same
ip addresses, but it's not guaranteed.

> A netmask might be workable if there were an entire class A
> network, properly routed and switched. From what I've seen though, the
> ISP's seem to generally stick with class C's.

Actually - subnetted Class C masks, usually for companies,
less for
individuals. It's hell getting larger assignments from an ISP pool.
Internet Address Authority regulations, not the ISP's.
You usually get blocks of 8, 16 or 32, maybe 64 host addresses if you are

> IIRC, one of the assigned netmasks had a 140 string, which
> doesn't appear
> to be a "legal" mask. This might be what threw Dr. Godwin off
> to suggest
> the mask.

sorry Sir Keith, Mr. Goodwin is NOT being thrown off by anything and where
you came up with the 140 above as part of his subnet masks is beyond me.
A typo perhaps, but that does not explain your subsequent comment about it's

Phil provided the following IP addresses
for his systems in his postings -

//Redhat: is a Class A network with a normal full net mask.

If I have computed everything correctly below........

His Ip addresses actually vary in the
lower 4 bits of the second octet, (238, 229) suggesting a POSSIBLE legal
subnet mask using the first 12 left most bits ( for the network
half of the subnet mask based on the 32 bit ip addresses. However even that
subnet mask scenario
would not be normal under his circumstances.

His masks provided are both perfectly legal

Valhalla mask: (according to ifconfig)
WinXP mask: (according to ipconfig /all)

Both are being applied against a Class A address with normally 8 bits of
network address
and 24 bits of host address.

The winxp is a normal 24 bit Class C full network mask giving him a network
of 255 host/broadcast addresses,
the valhalla mask (which might be the real netmask for his isp network
connection if he was a huge company)
is a normal 20 bit subnet mask.
It does not seem reasonable
in Phil's situation because it provides for 12 bits of host id and violates
IANA guidelines
by providing too many host addresses for the intended user.

Both address masks provide for FAR too many hosts for his setup. This leads
to believe that they are either irrelevant to his ISP (as Keith suggested)
or there is something else wrong on his network, like they never came from a
dhcp server in the first place. Phil should check with his cable company and
see what they have to say in regards
to the addresses and masks being provided.

The short of it is that none of the three possible subnet masks make sense
for his situation. They would therefore appear to be irrelevant and perhaps
replacable by the host mask of

> > Note also that this is probably the only way to get
> samba/smb to work
> > between the machines. (If you want to do this I cannot recommend
> > strongly enough placing the two machines behind an NAT firewall to
> > protect them from being hacked. This is inconvenient. But
> it will save
> > data for you.)
> NAT's only inconvenient until it's set up and running. After
> that, it's
> one of those things that you wish you had done long ago.
> Aside from smb,
> he can also run dhcp and a few other services not to be
> shared with the
> public.
> --
> Keith Mastin
> BeechTree Information Technology Services Inc.
> Toronto, Canada
> (416)696 6070
> _______________________________________________
> Valhalla-list mailing list
> Valhalla-list redhat com
> https://listman.redhat.com/mailman/listinfo/valhalla-list

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]