One Server -2 NICs, Different IPs
-
@frodooftheshire said in One Server -2 NICs, Different IPs:
When I did the initial onsite visit I thought the secondary connection was some sort of iLO or iDrac connection, but I can now see it seems like it's performing redundancy which isn't a huge concern as there are only maybe 6 users of this server. Like SAM said I should probably do some more digging before I disable it.
That is a possibility. I've seen people create a different IP for each NIC and then as "redundancy" call one A and one B in DNS. No idea what they are thinking, but I've seen it done.
-
@frodooftheshire said in One Server -2 NICs, Different IPs:
When I did the initial onsite visit I thought the secondary connection was some sort of iLO or iDrac connection (not sure if Lenovo has a version & what it's called), but I can now see it seems like it's performing redundancy which isn't a huge concern as there are only maybe 6 users of this server. Like SAM said I should probably do some more digging before I disable it.
Redundancy doesn't happen with a different IP address.
-
@scottalanmiller said in One Server -2 NICs, Different IPs:
@frodooftheshire said in One Server -2 NICs, Different IPs:
When I did the initial onsite visit I thought the secondary connection was some sort of iLO or iDrac connection (not sure if Lenovo has a version & what it's called), but I can now see it seems like it's performing redundancy which isn't a huge concern as there are only maybe 6 users of this server. Like SAM said I should probably do some more digging before I disable it.
Redundancy doesn't happen with a different IP address.
Correct. That's why it's in quotes. They called it redundancy and used computerB instead of computerA to connect...
-
@scottalanmiller Maybe I worded that wrong. I was just implying that because they had separate IPs they weren't utilizing the both nics in an effort to team them. In other words I had ruled out bandwidth aggregation, and figured they were trying to accomplish some sort of network redundancy.
-
@frodooftheshire said in One Server -2 NICs, Different IPs:
@scottalanmiller Maybe I worded that wrong. I was just implying that because they had separate IPs they weren't utilizing the both nics in an effort to team them. In other words I had ruled out bandwidth aggregation, and figured they were trying to accomplish some sort of network redundancy.
Reverse that logic. Since we can rule out redundancy, does that imply that they were trying to get better throughput?
-
lol, sounds like there was just a failure of understanding by the previous IT person.
I'm also wondering if it's virutalized, and if not, why not?
-
You could check for an IIS instance and see if it is listening on all IPs, or just a specific one. Then check DNS and see if there are entries for the second IP.
If it's a small network I would disconnect the extra NIC and listen for the screams. If you're remote when the call comes in, just add the second IP to the first NIC and that should get things going again.
-
Here's a thought - are the IPs on two different subnets?
-
Sometimes, the only way to find out what something does is to unplug it, then wait by the phone.
Haha, just kidding... but not really. Be careful!
-
@Dashrender The two IPs are on the same subnet. Yesterday when I was just gathering information remotely about different devices on the network via Advanced IP scanner that's when I saw same hostname/different IPs
When I tried to access 55.6 in a browser it loaded an IIS welcome page. That's the extent of what I know for now.
-
@Tim_G I could always disable the port on the switch and if the screams come in switch it back on. Like most clients I take on the customer has zero documentation on how things are setup/configured. Maybe I'll be able to have a conversation with the old IT guy and see if he's willing to share on why things the way they are.
-
Open up resmon.exe, network tab.
TCP Connections section, then look at Local Address for the IPs, and see what's connecting to them.
-
It very well could be that the server had several network ports, and someone just figured they all aught to all be plugged in... just out of ignorance. I'd see what traffic is on them as I said above. Or what you said, just ask them why.
-
@Tim_G That's a good idea. I'm pretty sure it was ignorance. I mean...they have a Windows server and installed the DHCP role, but did not configure it. DNS is somewhat configured correctly but all the endpoints are pointing to Comcast's DNS servers...soooo.
-
@frodooftheshire said in One Server -2 NICs, Different IPs:
but all the endpoints are pointing to Comcast's DNS servers
I'm sooo sorry.
-
As much as I am against downtime, I might unplug the second NIC. If there is an issue, it just needs to be plugged back in. If there isn't an issue for a month, you can team the NICS.
If you don't really see any traffic on the second NIC. So what is the actual risk? perhaps downtime, but all you need to do is plug it back in if there is an issue.
-
At the end of the day, you can spend hours trying to figure this out or you can save the customer money and find out if it is needed prettty easily.
-
@IRJ said in One Server -2 NICs, Different IPs:
At the end of the day, you can spend hours trying to figure this out or you can save the customer money and find out if it is needed prettty easily.
Exactly - Though just a CYA, tell the management you are doing that, and let them say it's OK to save the money that way...
-
If the IPs are on the same subnet, it's doing more harm than good. Traffic coming in on one NIC will go out the other. Windows likes to use the lowest IP address to send data out. If you receive on 56, it will go out on 55. I've seen this happen. We have wireshark logs.
I would suggest you immediately remove or team the spare nic to increase throughput. Be sure that your network hardware supports that!
You can do a route print from the command line to see which nic is dominant.
@frodooftheshire said in One Server -2 NICs, Different IPs:
@Tim_G I could always disable the port on the switch and if the screams come in switch it back on. Like most clients I take on the customer has zero documentation on how things are setup/configured. Maybe I'll be able to have a conversation with the old IT guy and see if he's willing to share on why things the way they are.
I'm working in an environment now with this kind of bullshit and it's almost 100% because the previous admins were /idiots./
-
If you can find a non-critical time frame it might well be worth just shutting one of them off and seeing what breaks. If nothing, leave it off.
