Machines returning same IP address

[tomscaper]

Not sure how to explain this one, but we have about 250 machines on the network, but there are about 4 or 5 machines that when i try to ping them by the computer name they return the same IP address.

 

Hopefully this is something easy to resolve.

 

Have attached screen shot of two computers i know about.

Edited April 30, 2008 by tomscaper

[Jona]

Hmmm are these machines reporting IP conflicts at all? what happens if you do nslookup for both machines?

[Geoff]

Enable scavanging on your rDNS zone.

[tomscaper]

I'd check the DNS entries on the server. I take it they have different IPs when you do ipconfig on the machines locally.

 

Yes if i were to do ipconfig on the machine themselves it would bring up different IPs

 

I am not sure how i would check the DNS entries on the server, any chance you could explain where to look.

 

The computers themselves all work and connect, but for some reason i cant ping them by name.

[tomscaper]

This is what nslookup reports

 

Hmmm are these machines reporting IP conflicts at all? what happens if you do nslookup for both machines?

[Geoff]

In the above examples, what are those two machines real IP addresses?

[tomscaper]

staff-01 has 10.64.17.176

res-13 has 10.64.17.68

In the above examples, what are those two machines real IP addresses?

[Guest blacksheep]

I would firstly flush the local dns cache (ipconfig /flushdns)

 

Then repeat what youve done. If it the same then goto the DNS management console on your server or wks (administrative tools->DNS) Look at your forward lookup zones, in the domain name you should see a list of A records. Look for the entry and see if they have repeats, you can delete them here. Then do an ipconfig /registerdns and they should reappear in the dns properly. If someone has added them manually there could also be a C Name record for example, just check the names.

If this has happened then for it to work proper in the future the DNS scavenge needs to be switched on and setup with the right settings (as well as dhcp). Actually that reminds me about mine

[Sirbendy]

yep. IPCONFIG /FLUSHDNS.

 

I end up doing it here sometimes..Netsupport does something similar on rare occasions..heh.

[FatBoy]

We sometimes get duplicate entries in our DNS I guess because the machine aren't sometimes renamed or removed from the domain properly just delete the bad entires in DNS

 

@ Geoff how would you enable scavanging on your rDNS zone and should this be something thats done on a LAN?

[torledo]

We sometimes get duplicate entries in our DNS I guess because the machine aren't sometimes renamed or removed from the domain properly just delete the bad entires in DNS

 

@ Geoff how would you enable scavanging on your rDNS zone and should this be something thats done on a LAN?

 

right click on the zone, and there's an option to set scavenging time or something like that. I believe the default is 14 days but You can then set the aging time in days to whatever you wish. There's also an option to scavenge resource records now, although can't remember if that's under rDNS options or somewhere else.

 

We set scavenging to seven days and tried the scavenge now option...neither worked, so i think i'll have to just delete the duplicate records manually.

[m25man]

Your scavenging period should be set to the duration of your DHCP leases.

It's a common mistake to change DHCP lease durations without changing the DNS scavenging periods.

This results in discrepancies in your DNS/DHCP databases with machines appearing to have the same or duplicate IP addresses in the DNS zone.

Whilst this does not normally affect the client PC's themselves if they share files or printers, others will not be able to find them and Group Policies will regularly fail to apply on machines with multiple reverse DNS entries.

 

Others have already correctly advised you how to set this.

 

DNS DNS DNS, this is the core of Windows AD get this right and the rest is like falling off a log!

Get it wrong and your AD will eventually die, along with your reputation.

[FatBoy]

Ours is set to 8 days but I think it's got a lot to d with how many clients you have and what sort of DHCP range size you have set.

 

Also FYI I was looking around some of are settings and if you right click on your Scope in DHCP and go to properties, then the DNS tab and under here you have "Enable DNS dynamic updates according to the setting below". Ours was set to only update if the clients requests it so I have now changed it to "Always dynamically update DNS A nad PTR records"

 

[tmcd35]

Your scavenging period should be set to the duration of your DHCP leases.

It's a common mistake to change DHCP lease durations without changing the DNS scavenging periods.

This results in discrepancies in your DNS/DHCP databases with machines appearing to have the same or duplicate IP addresses in the DNS zone.

Whilst this does not normally affect the client PC's themselves if they share files or printers, others will not be able to find them and Group Policies will regularly fail to apply on machines with multiple reverse DNS entries.

 

Others have already correctly advised you how to set this.

 

DNS DNS DNS, this is the core of Windows AD get this right and the rest is like falling off a log!

Get it wrong and your AD will eventually die, along with your reputation.

 

Have some rep! just checked my settings, way off - explains the odd discrepancy I've seen bit like the OP. All changed now

[Guest blacksheep]

Read some time back -

 

The no refresh interval is the time that client cannot re-register itself. This to stop machines overly re-registering and potentially causing lots of replication traffic between DCs (mainly upsetting WAN etc.)

 

The refresh interval - the time the record is immune from being scavenged

 

If the records age is less than the sum of the refresh and no-refresh then it will never be scavenged/deleted.

 

The frequency of these should be larger than how often a new ddns registeration takes place, so as a registeration takes place on a dhcp renewal / when you turn you computer on (prob daily except hols) and I think every 24hrs if the machine is on.

 

Mine is set to 7 and 7, so longest is 14 days. DHCP is on 7 days which seems to work ok here.

 

So if all the machines were off for more than 2 weeks they could lose their DNS entry, but this doesnt cause probs for me as important stuff like servers etc. are set statically anyway.

 

hope this clarifys it somewhat!

Edited May 1, 2008 by blacksheep

[torledo]

Read some time back -

 

The no refresh interval is the time that client cannot re-register itself. This to stop machines overly re-registering and potentially causing lots of replication traffic between DCs (mainly upsetting WAN etc.)

 

The refresh interval - the time the record is immune from being scavenged

 

If the records age is less than the sum of the refresh and no-refresh then it will never be scavenged/deleted.

 

The frequency of these should be larger than how often a new ddns registeration takes place, so as a registeration takes place on a dhcp renewal / when you turn you computer on (prob daily except hols) and I think every 24hrs if the machine is on.

 

Mine is set to 7 and 7, so longest is 14 days. DHCP is on 7 days which seems to work ok here.

 

So if all the machines were off for more than 2 weeks they could lose their DNS entry, but this doesnt cause probs for me as important stuff like servers etc. are set statically anyway.

 

hope this clarifys it somewhat!

 

thanks for in the info, but i read today that the refresh interval and the no-refresh interval combined should not exceed the dhcp lease duration....

 

so here we have 8 day dhcp lease duration, and i've set the scavenging period to 8 in dns, the refresh to 5 and the no-fresh to 3 for both forward and rdns zones. I don't know why i didn't do 4 and 4 but i'm told 5 and 3 should be fine.