Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

We have a existing class-c network with IP address range 11.22.33.44/24 (just for example). My domain controller has been configured within this subnet. So all servers within this subnet have subnet mask configured to 255.255.255.0.

Now we have got a new subnet with IP address 11.22.88.99/24 (note that only last 2 octets have changed). I want all new hosts in this new subnet to join my existing DC. For this we have configured firewall properly so allow this. (so there is no issue with firewall).

But initially I was not able to join hosts in new subnet in existing domain. Later I doubted on subnet mask used in domain controller (255.255.255.0) and for testing purpose I changed it to 255.255.0.0, it worked like charm, i was able to join subnet-2 hosts in subnet-1 domain.

Now i am wondering whether it will be good practice to change subnet mask of a class-c network to 255.255.0.0? Can any issues arise due to this? Experts please provide your opinion.

share|improve this question
3  
"Class C" is archaic terminology. It's pronounced "/24" today. –  Evan Anderson May 26 '10 at 13:35
    
I guess I mis-interpreted my network as class C network. All hosts in my subnet have public IP addresses starting with 87.233.x.y (no NATing, public IP assgined directly to host NIC). Which means my network is just a small subnet from a big class-A network. Experts, please correct me if I am wrong –  TechCoze May 26 '10 at 13:48
    
So you're network is 87.233.a.b, and the new one is 87.233.x.y; If a and x are different numbers more than 1 different, you'll end up with problems changing the network mask. –  Chris S May 26 '10 at 13:53
    
@Prashant Mandhare: The terminology "Class A", "Class C", etc, is archaic terminology. Classless inter-domain routing (CIDR - faqs.org/rfcs/rfc4632.html) obseletes that terminology. I'm, perhaps, being pedantic about terminology, but the "classes" of IP addresses are wholly irrelevant in modern IPv4 routing. –  Evan Anderson May 26 '10 at 13:59
1  
Assuming you are using 87.233.x.y /24 and have a new net 87.233.a.b /24 then joining them with 87.233.0.0 /16 is not the answer as was pointed out below. Are all of the devices that might use those IP's in the same collision domain? Assuming you have a legitimate need for each device to be internet visible, I would reassign private IP's to all of the devices and then Static NAT them to the public IP's. –  dbasnett May 26 '10 at 14:00
add comment

2 Answers

up vote 2 down vote accepted

The issue is that your DC did not know how to route to the new subnet. In the case of Internet routable addressing, changing your subnet mask to be substantially larger without considering the full implications of your network could potentially introduce routing issues to hosts within that subnet.

Add a static route to the DC for 11.22.88.99/24. Depending on your infrastructure, you may be able to just add it to the firewall or router between the subnets. If you continue to have issues, your network administrator would be able to identify the proper route.

share|improve this answer
1  
I was going to downvote, but I'm going to stay my hand in light of the poster not giving us enough information. Nonetheless, there is nothing illegal or bothersome about using a "/16" on a LAN subnet. Certainly, if there are other subnets between the current "/24" and the new "/24" in use in other areas of his network then using a "/16" that overlaps both would be a bad idea. In principle, though, you can use any subnet mask you want (so long as it has enough available host IDs) on a LAN and have no problems. Now, having too many actual hosts for the network medium-- that's actually a problem. –  Evan Anderson May 26 '10 at 13:41
1  
Thank you for the stay of hand. At this point, I'd been considering deleting my answer but I want to respond to your comment. Certainly, I was making an assumption regarding intranet routable subnets. I'm surprised he's working with Internet routable addresses. While there's nothing bothersome with using a /16, arbitrarily changing the subnet on a specific box has implications beyond the understanding exemplified in the question. Often it doesn't seem reasonable to write a 2-page essay to a question with limited distinction. –  Warner May 26 '10 at 14:39
    
One downside to using big subnets down the road is if your company ever peers with other companies. So for example, if you have three offices and make them 10/8, 172.16/12, and 192.168/16 then all of the sudden you merge with another company, or need VPN to their datacenter, you will have a collision. Of course you might be able to work around it with some nat, rearranging, and could have a collision anyways. However, I think better to be at least mildly conservative. –  Kyle Brandt May 27 '10 at 17:58
add comment

If you're using public address like the ones in your example, yes. There's a few thousand websites your computers will no be able to route to because they'll think the destination is on the local network.

If you're actually using private IPs, you'll likely run into many other problems as time goes on.

If the network numbers are not contiguous, you should leave them on separate subnets and use a router to connect them. If they are contiguous, you need to figure out the smallest netmask that makes everything work (or just go with the router).

share|improve this answer
    
Using a "/16" for a small private LAN subnet seems wasteful, but there isn't anything inherently wrong with it. The poster saying "we got a new subnet" makes me think that there's an IT group in his company "assigning" subnets, in which case changing subnet masks on existing server computers is probably a bad idea. The poster should clarify why he's being forced to use this other subnet to add some helpful background. –  Evan Anderson May 26 '10 at 13:38
1  
Our servers're in a third party data center & they assign us subnets, we don't have control over subnet config. Now in beginning they assigned us subnet with 87.233.a.b/24. now when we were running short of IP addresses in first subnet, we've been provided with new subnet 87.233.x.y from data center (diff between a and x is > 1). So it's obvious that intermediate n/w ranges have been assigned to other clients. Data center technician told us that a routing link has been established between old & new subnet, but still we were facing domain joining issue, hence we thought of mask change –  TechCoze May 26 '10 at 14:09
1  
Sounds like your default router isn't the one they added the route for the new subnet to. Just need to figure out what router they added it to, and add the correct route to your default router. –  Chris S May 26 '10 at 14:18
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.