Morning,

Grateful for any help, my unix knowledge is very old and rusty, coming
from a predominantly VMS-background ...

Scenario is this -

Site A is head office, in London, with a fileserver (win 2003),
exchange server (win 2003), and an AIX box. All can talk to each other
quite happily, ping responses are fast and telnet access is fine.

Site B is located in Birmingham, and has access to head office via a
secure VPN.

Site C is located in Bristol, and also has access to head office via a
secure VPN.

Prior to tuesday this week, everything was fine, the remote users
could access their email, see the shared drives on the fileserver, and
telnet to the unix box ... however, as of late tuesday afternoon Site
B can no longer telnet to the unix system, ping does not respond, and
a traceroute dies after entering the local lan at head office.

So, in short -

Site A is fine

Site B can access email, fileserver, but cannot telnet to unix box,
however they can telnet to other units in Site A so I know its not
port 23 being blocked.

Site C is fine

The VPN from Site B to Site A is fine, from Site B I can ping,
traceroute and telnet to the exchange and fileservers with a very fast
response but pinging the unix box gives request timeouts ... pinging,
traceroute and telnet from Site C is also fine.

I found a note on another newsgroup regarding tcp streams, which
sounded very similar to the problem here, but as I mentioned earlier
my AIX knowledge is slim to non-existant, so any pointers would be
very appreciated, if there is anyone who can help - if you need any
further info please do ask.

Many thanks to anyone who can shed some light on this !

Cheers,
Jeff.

In article <8e97f4ef.0405260032.42e544fe@posting.google.com >, Jeff wrote:
> Morning,
>
> Grateful for any help, my unix knowledge is very old and rusty, coming
> from a predominantly VMS-background ...
>
> Scenario is this -
>
> Site A is head office, in London, with a fileserver (win 2003),
> exchange server (win 2003), and an AIX box. All can talk to each other
> quite happily, ping responses are fast and telnet access is fine.
>
> Site B is located in Birmingham, and has access to head office via a
> secure VPN.
>
> Site C is located in Bristol, and also has access to head office via a
> secure VPN.
>
> Prior to tuesday this week, everything was fine, the remote users
> could access their email, see the shared drives on the fileserver, and
> telnet to the unix box ... however, as of late tuesday afternoon Site
> B can no longer telnet to the unix system, ping does not respond, and
> a traceroute dies after entering the local lan at head office.
>
> So, in short -
>
> Site A is fine
>
> Site B can access email, fileserver, but cannot telnet to unix box,
> however they can telnet to other units in Site A so I know its not
> port 23 being blocked.
>
> Site C is fine
>
> The VPN from Site B to Site A is fine, from Site B I can ping,
> traceroute and telnet to the exchange and fileservers with a very fast
> response but pinging the unix box gives request timeouts ... pinging,
> traceroute and telnet from Site C is also fine.
>
> I found a note on another newsgroup regarding tcp streams, which
> sounded very similar to the problem here, but as I mentioned earlier
> my AIX knowledge is slim to non-existant, so any pointers would be
> very appreciated, if there is anyone who can help - if you need any
> further info please do ask.
>
> Many thanks to anyone who can shed some light on this !
>
> Cheers,
> Jeff.

When you're unable to telnet, what error do you get? Did the host
at A have a change either to its default route, or some route such
that it can't see backwards to B (can A ping B?), or on some
hosts the file /etc/hosts.allow might disallow sites from talking.

Mike

Mike <mikee@mikee.ath.cx> wrote in message news:<10b9gurtof7mp14@corp.supernews.com>...
>
> When you're unable to telnet, what error do you get? Did the host
> at A have a change either to its default route, or some route such
> that it can't see backwards to B (can A ping B?), or on some
> hosts the file /etc/hosts.allow might disallow sites from talking.
>
> Mike

Hi Mike,

Thanks for the response ... I found the problem, not quite sure what
happened though - the routing table entries on the AIX box for site B
had disappeared ... reason why that is odd is the only person who
would have any clue how to remove them is me, and obviously I did not
do it lol ... only other explanation I can think of is some sort of
table corruption but I would have expected the lot to go if that
happened, dunno whether its possible for a single route to corrupt ?

Anyway all is working fine now, many thanks for the pointer.

Cheers,
Jeff.

.... I found the problem, not quite sure what
> happened though - the routing table entries on the AIX box for site B
> had disappeared ...

This may have to do with how the routes were added. I noticed similar
behavior on some new systems lately, and when I added the routes via
smit, they stayed around. Routes added interactively, disappeared.
In some cases I know, elements edited/added via smit are put into the
ODM, whereas cmd line edits/additions are not.

Matt