Inproper subnet mask

[Nuwave]

Hello all,

I have come across a bunch of PC's in a corporate environment in a subnet with the wrong subnet mask. The mask they are supposed to have is a /22 but they were found with a /24.

These computers are in a newly constructed building. With 2 layer 2 switch stacks, and 1 layer 3 switch for connectivity back to the MAN. The default gateway has the same first three octets for an IP eg a machine would have 10.1.1.1 as an IP and the gateway would be 10.1.1.200.

Long story short there was 'weired' problems occuring there, I was wondering from all you other professionals what problems can occur, or have seen on your travels with improper subnet masks on the end machine.

Thanks in advance for your responces.

 

[spidey07]

What happens all depends on the configuration of the default gateway router and other routers attached to that broadcast domain. If they tried to reach addresses in the other 3 class C networks the default gateway would most likely send a ICMP redirect to the sending host, it may even just forward it.

If they didn't have the correct default gateway listed then the router would have to do proxy arp for them to talk to anything outside their own /24 network. And we hate proxy arp so we normally disable it.

The return traffic wouldn't matter much as long as the layer2 and layer3 addresses were correct on the frame/packet respectively. It could also cause weird behavior with protocols that rely on broadcast like netbios/tcp name resolution.

 

[Nuwave]

I was seeing packetloss from devices from (and including) one of the switch stacks.

Ping successfull
Ping successfull
Ping failed
Ping failed
Ping successfull
Ping successfull
Ping successfull
Ping failed
etc....

All the devices on the other switch stack where fine, no packetloss. So I figured there was a cabling issue / switch port issue or something wrong with the switch itself. I ended up replacing the switch which seemed to fix the problem. The machine in question started working beautifully and the user was extremely happy. We are a citrix environment and delays in the network cause any application to have typing delays.

However all the devices on the other switch stack starting having packet loss like above after the booting of the 1st switch stack. To add to the confusion, while the PC's were not working worth 2 cents the VoIP phones were working just fine. I knew they were on a tagged vlan and have adifferant Gateway. But it wasn't making sence at first, I know now that it is most likely caused by the Improper subnet mask. I then rebooted the 2nd switch stack in hopes that I could get the network fuctionality back to the way it was before I did anything.

The thing I'm still confused about is why the symptoms only seemed to effect devices off of one switch stack at a time, and rebooting a switch stack seemed to force the problem on to the other.