Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
para
Nov 30, 2006

Tony Montana posted:

So I wrote ACL 180:

Extended IP access list 180
10 permit tcp any any eq smtp

I then ran the ip packet debug with the ACL as a modification:

debug ip packet 180 detail

This is a neat idea but wouldn't it essentially severe your connection because of the implicit deny all?

You could install some sort of netflow solution to help you see all this traffic without needing to disrupt normal operations. You didn't say how big your network was, but ManageEngine NetFlow Analyzer has a free 30 day trial, after which it is limited to two interfaces. Maybe that would help?

Adbot
ADBOT LOVES YOU

nex
Jul 23, 2001

øæå¨æøåø
Grimey Drawer
What is your routine when upgrading remote critical routers?

Personally I triple check the IOS image checksum, verify bootvar and all that. I also dump out BGP and multicast summaries so that I can verify that the routers come up in the same state that they went down in.

Any other neat tricks and quirks that might come in handy?

jwh
Jun 12, 2002

routenull0 posted:

You have to put the ACL in the proper direction on the interface. Inbound or Outbound. You have it set to inbound, meaning traffic coming from the internet, not in from your lan.

Actually, debug ip pack doesn't need an access-list to be applied to an interface, but it does need the traffic to be process switched. I'm wondering if that's part of the problem here, although NAT is in the CEF path now, so I don't know why the inbound flows would be process switched but the outbound flows wouldn't.

Personally, I think the easiest way to look at this would be with 'show ip nat translation'

Tony Montana
Aug 6, 2005

by FactsAreUseless

routenull0 posted:

You have to put the ACL in the proper direction on the interface. Inbound or Outbound. You have it set to inbound, meaning traffic coming from the internet, not in from your lan.

para posted:

This is a neat idea but wouldn't it essentially severe your connection because of the implicit deny all?

The ACL isn't bound to the interface. It's there just for use with the debug.

jwh posted:

Actually, debug ip pack doesn't need an access-list to be applied to an interface, but it does need the traffic to be process switched. I'm wondering if that's part of the problem here, although NAT is in the CEF path now, so I don't know why the inbound flows would be process switched but the outbound flows wouldn't.

Personally, I think the easiest way to look at this would be with 'show ip nat translation'

Jesus jwh, now I need to go and look up 'process switched' in a Cisco context. You should really get some certs, you're easily as gun as many of the TAC engineers I've worked with. I'm going to try 'ip nat translation', see how I go.

As for extra monitoring software, I've got a couple that I am familiar with anyways. It's more trying to get IOS to do it, more of an exercise for aspiring Cisco techs as I think mastering debug outputs puts you in a much better position to understand what is happening.

Here is something else, how do people feel about the term 'engineer'? I personally have a bit of a problem with it, I just don't feel I've done the heavy math and science required to call myself an engineer. I know even Cisco calls their technical employees 'engineers', I had a contractor called a 'network architect' the other day.

It just makes me feel uncomfortable, I dread running into someone that says 'engineer? oh cool, I graduated from Adelaide Uni, where did you?'.

falz
Jan 29, 2005

01100110 01100001 01101100 01111010

nex posted:

What is your routine when upgrading remote critical routers?

Personally I triple check the IOS image checksum, verify bootvar and all that. I also dump out BGP and multicast summaries so that I can verify that the routers come up in the same state that they went down in.

Any other neat tricks and quirks that might come in handy?
Modem connected to it and test the new image/config in Dynamips are the only two things that come to mind.

jwh
Jun 12, 2002

Tony Montana posted:

Jesus jwh, now I need to go and look up 'process switched' in a Cisco context. You should really get some certs, you're easily as gun as many of the TAC engineers I've worked with. I'm going to try 'ip nat translation', see how I go.
Thanks for the compliment. There are a couple of different forwarding paths in IOS, the slowest of which is "process switching". There's also Fast Switching, which you hardly ever see anymore, and CEF which is Cisco Express Forwarding. Or Customer Enragement Feature, depending on whether you have an axe to grind. Long story short, there are different forwarding paths, and they involve either switching at an interrupt level or placing particle data in system memory and calling IP_Input to resolve a forwarding adjacency. You can read about it if you're really interested.

Tony Montana posted:

Here is something else, how do people feel about the term 'engineer'?

I like it more than analyst or administrator, because when I think of an engineer, I think of someone building something. Or running around with explosives and a bolt action rifle, but I blame that on Battlefield 1942. Either way, there's no Analyst class in Battlefield 1942, so I think engineer wins.

I once did get chastised however for referring to myself as an engineer- some guy took offense to that, and claimed that if I hadn't passed the state engineering exam, I wasn't "legally allowed to call myself an engineer." However that guy was a jerk, so whatever.

jwh fucked around with this message at 05:44 on May 26, 2009

cptInsane0
Apr 11, 2007

...and a clown with no head
I agree with JWH. Also I majored in Computer Engineering, and HAVE taken all of the math and science classes. Network Admins are not the same thing at all.

inignot
Sep 1, 2003

WWBCD?

jwh posted:

I once did get chastised however for referring to myself as an engineer- some guy took offense to that, and claimed that if I hadn't passed the state engineering exam, I wasn't "legally allowed to call myself an engineer." However that guy was a jerk, so whatever.

Should have asked him if he had passed the bar exam, otherwise he can't legally tell you what you're legally allowed to be called.

cptInsane0
Apr 11, 2007

...and a clown with no head
nice. I think his head would explode.

Tony Montana
Aug 6, 2005

by FactsAreUseless

cptInsane0 posted:

I agree with JWH. Also I majored in Computer Engineering, and HAVE taken all of the math and science classes. Network Admins are not the same thing at all.

But you're not agreeing with jwh. He's saying he doesn't have a problem with calling himself an engineer, and once someone picked him up on it but he didn't care because they were silly anyway.

I think you're more agreeing with me, that because I haven't done actual engineering at a university it annoys you (that has done it) because I have some Cisco certs and a stack of experience I call myself an engineer. I do hate the 'administrator' term though.

You all are probably thinking I'm obsessing over this, but it's important to me.

jwh
Jun 12, 2002

Well, I tend to think of it like this: an engineer builds networks, an administrator takes care of them. I'm being too literal, probably.

Tony Montana
Aug 6, 2005

by FactsAreUseless
Ok, well I resolved my debugging problem. Looks like jwh was right on the money, but there are a couple of fundamental misunderstandings that needed to be worked through.

I'll attach my notes here incase someone is interested. Feel free to correct me, I've drawn some conclusions that might not be completely right.

me posted:

Ok, so the intention here is to run a packet debug but restrict the output make it more useful. Something like 'show all SMTP packets going into an interface'. The 'debug ip packet' command will show all ip packets going through the router, however this output is massive and will be hard to sift through to do meaningful troubleshooting. So what we want to do is write an access list defining the output we want to see, then run the 'ip packet' debug filtered through that access list.

There was some loving around to get this going!

1) The first concept to understand is 'ip route-caching'. This is how the router reduces it's CPU workload, when a packet comes in and the router looks it up in the routing table, it makes a choice and forwards the packet. That choice is stored in cache memory. So the next packet that comes along with the same source and destination, the router doesn't have to look it up again, it just calls cache and uses that without bothering the CPU. HOWEVER if the router just uses cache the packet has not been 'packet-switched' (meaning the router didn't look it up in the routing table) and WON'T BE VISIBLE IN ANY DEBUG because debugs just show packet-switched packets. This means you need to go to the outgoing interface (usually Dialer0 on a 800 series router) and the inbound interface (usually VLAN1 or BVI1) and disable this caching (the command to do this is 'no ip route-cache'). This is a BAD thing though, because the router now needs to process all these packets it didn't need to previously, driving up CPU use. This is the first of a couple of things you'll want to change for the troubleshooting process ONLY, then put it back the way it was before!

2) The next concept is writing the ACL and how the matching works. Firstly write the ACL that'll modify the debug output, usually a 'permit any any' and specifing the port. You don't bind this ACL to any interface, it'll only be used in the debug. However a trick is you can only match a single ACL entry on an interface, meaning if you write two ACLs both allowing SMTP traffic and bind them to an interface only one will be used. You can see which by doing a 'show access-lists' and seeing the matches. So the problem here is that if you write a 'permit any any eq 25' for your debug, but there is already the same line in the ACL actually applied to the interface the 'show access-lists' will show you that your packets are matching the interface ACL, not your debug. So to work around this I took the SMTP line out of the interface ACL, put in a 'permit any any' because otherwise SMTP packets would now be denied and then the debug worked, I could see the matches matching to the debug ACL. Asssuming your debug ACL is numbered '180' the line to display customised debug would be 'debug ip packet detail 180'. Remember you've opened the router right up with your 'permit any any' on the incoming interface! This is the second thing you want for the troubleshoot process only, put it back after you've finished!

That's it! Here is an example of some of the output you're trying to get. In this example 10.1.8.1 is the client's mail server, 210.65.130.15 is our mail server and 219.26.189.25 is the NATted WAN IP of the client's mail server. Notice the first line shows you the LAN source being forwarded to the WAN IP of the destination. Then you can see the packets being sent to our mail server, there are a few because I think these are the indivial SMTP commands being executed (HELO, MAIL TO, MAIL FROM, DATA) you can see the DATA command bigger than the rest (because it's actually containing the data of the email) with the 'len 1258'.

May 27 15:15:06.227: IP: tableid=0, s=10.1.8.1 (BVI1), d=210.65.130.15 (Dialer0), routed via FIB
May 27 15:15:06.227: IP: s=219.26.189.25 (BVI1), d=210.65.130.15 (Dialer0), g=210.65.130.15, len 48, forward
May 27 15:15:06.227: TCP src=48228, dst=25, seq=1989788012, ack=0, win=65535 SYN
May 27 15:15:06.283: IP: tableid=0, s=10.1.8.1 (BVI1), d=210.65.130.15 (Dialer0), routed via FIB
May 27 15:15:06.283: IP: s=219.26.189.25 (BVI1), d=210.65.130.15 (Dialer0), g=210.65.130.15, len 40, forward
May 27 15:15:06.283: TCP src=48228, dst=25, seq=1989788013, ack=628415245, win=65535 ACK
May 27 15:15:06.339: IP: tableid=0, s=10.1.8.1 (BVI1), d=210.65.130.15 (Dialer0), routed via FIB
May 27 15:15:06.339: IP: s=219.26.189.25 (BVI1), d=210.65.130.15 (Dialer0), g=210.65.130.15, len 62, forward
May 27 15:15:06.339: TCP src=48228, dst=25, seq=1989788013, ack=628415345, win=65435 ACK PSH
May 27 15:15:06.435: IP: tableid=0, s=10.1.8.1 (BVI1), d=210.65.130.15 (Dialer0), routed via FIB
May 27 15:15:06.435: IP: s=219.26.189.25 (BVI1), d=210.65.130.15 (Dialer0), g=210.65.130.15, len 90, forward
May 27 15:15:06.435: TCP src=48228, dst=25, seq=1989788035, ack=628415536, win=65244 ACK PSH
May 27 15:15:06.515: IP: tableid=0, s=10.1.8.1 (BVI1), d=210.65.130.15 (Dialer0), routed via FIB
May 27 15:15:06.515: IP: s=219.26.189.25 (BVI1), d=210.65.130.15 (Dialer0), g=210.65.130.15, len 75, forward
May 27 15:15:06.515: TCP src=48228, dst=25, seq=1989788085, ack=628415557, win=65223 ACK PSH
May 27 15:15:06.579: IP: tableid=0, s=10.1.8.1 (BVI1), d=210.65.130.15 (Dialer0), routed via FIB
May 27 15:15:06.579: IP: s=219.26.189.25 (BVI1), d=210.65.130.15 (Dialer0), g=210.65.130.15, len 56, forward
May 27 15:15:06.583: TCP src=48228, dst=25, seq=1989788120, ack=628415581, win=65199 ACK PSH
May 27 15:15:06.827: IP: tableid=0, s=10.1.8.1 (BVI1), d=210.65.130.15 (Dialer0), routed via FIB
May 27 15:15:06.827: IP: s=219.26.189.25 (BVI1), d=210.65.130.15 (Dialer0), g=210.65.130.15, len 1258, forward
May 27 15:15:06.827: TCP src=48228, dst=25, seq=1989788136, ack=628415581, win=65199 ACK PSH
May 27 15:15:07.203: IP: tableid=0, s=10.1.8.1 (BVI1), d=210.65.130.15 (Dialer0), routed via FIB
May 27 15:15:07.203: IP: s=219.26.189.25 (BVI1), d=210.65.130.15 (Dialer0), g=210.65.130.15, len 46, forward
May 27 15:15:07.203: TCP src=48228, dst=25, seq=1989789354, ack=628415680, win=65100 ACK PSH
May 27 15:15:07.259: IP: tableid=0, s=10.1.8.1 (BVI1), d=210.65.130.15 (Dialer0), routed via FIB
May 27 15:15:07.259: IP: s=219.26.189.25 (BVI1), d=210.65.130.15 (Dialer0), g=210.65.130.15, len 40, forward
May 27 15:15:07.259: TCP src=48228, dst=25, seq=1989789360, ack=628415728, win=65052 ACK FIN
May 27 15:15:07.259: IP: tableid=0, s=10.1.8.1 (BVI1), d=210.65.130.15 (Dialer0), routed via FIB
May 27 15:15:07.259: IP: s=219.26.189.25 (BVI1), d=210.65.130.15 (Dialer0), g=210.65.130.15, len 40, forward
May 27 15:15:07.259: TCP src=48228, dst=25, seq=1989789361, ack=628415729, win=65052 ACK

For my little exercise, any SMTP packet originating from inside the LAN that isn't the mail server (10.1.8.1) and you've found your mailing worm!

cptInsane0
Apr 11, 2007

...and a clown with no head
I agree more with JTW. I did go to college for engineering, but that's not what makes me a network engineer. I could have gotten this job without the college.

H.R. Paperstacks
May 1, 2006

This is America
My president is black
and my Lambo is blue
Paycheck is what counts, not the title.

I'll engineer networks even if my title is "poo poo hauler" as long as the paycheck cashes for the correct amount.

H.R. Paperstacks fucked around with this message at 16:51 on May 27, 2009

ate shit on live tv
Feb 15, 2004

by Azathoth
Engineer is a more prestigious title than specialist, or analyst or whatever, but it has gotten diluted over the years. So who cares really.

cptInsane0
Apr 11, 2007

...and a clown with no head

Powercrazy posted:

Engineer is a more prestigious title than specialist, or analyst or whatever, but it has gotten diluted over the years. So who cares really.

Agreed.

jwh
Jun 12, 2002

Latest 3560 code, 12.2(50)SE1 has a weird bug where you can't log into the switch for about three or so minutes after it boots. Pressing enter either at console or via vty results in the switch throwing an authorization denied with no username visible. After a certain amount of time, it suddenly acts "normal" and presents the Username: prompt.

I wonder how that one got by QA.

skipdogg
Nov 29, 2004
Resident SRT-4 Expert

Quick question about OSPF.

Is there a way to delete just one OSPF route/process instead of reloading the whole thing?

I'm not the network person, but we replaced a firewall today, and it wouldn't see it until we actually ran a clear ip ospf and let everything rebuild which caused a blip in connectivity.

jwh
Jun 12, 2002

What wouldn't see the new firewall? The OSPF process on a neighboring router?

ragzilla
Sep 9, 2005
don't ask me, i only work here


skipdogg posted:

Quick question about OSPF.

Is there a way to delete just one OSPF route/process instead of reloading the whole thing?

I'm not the network person, but we replaced a firewall today, and it wouldn't see it until we actually ran a clear ip ospf and let everything rebuild which caused a blip in connectivity.

You should have a "clear ip ospf process <process id>" command to clear a specific process.

skipdogg
Nov 29, 2004
Resident SRT-4 Expert

jwh posted:

What wouldn't see the new firewall? The OSPF process on a neighboring router?

We switched from our normal netscreen 204 to our backup netscreen 204. cisco gear in question is a 6513 running IOS acting as a core router/switch.

We replaced each connection and the drat thing would work, couldn't ping the firewall interface or anything. It was strange, but I'm not the network person. I thought it might have something do with with the ARP cache but clearing that didn't help. The backup firewall originally had a different IP address when we configured it, we switched it to the same IP as the 1st one, but the OSPF route kept referencing the old .100 ip when it should have been .21. It was strange, clearing ospf was a last ditch thing, but it worked.

some kinda jackal
Feb 25, 2003

 
 

routenull0 posted:

Paycheck is what counts, not the title.

I'll engineer networks even if my title is "poo poo hauler" as long as the paycheck cashes for the correct amount.

Truth. I was "Sr. Software Engineer" at the last place I worked and the way they cobbled that title together was the most ridiculous thing seeing as how I had less than a year experience programming at that time, surrounded by people with lesser titles who had been doing it most of their life. For what it's worth I told them I didn't feel right taking the title but they said it was linked to the pay grade so my arguments promptly stopped :)

Beurocracy! :effort:

para
Nov 30, 2006
Let's say you have two access points in bridge mode, one of them was set as the root and the other a non-root. There is very little interference in the air, but they are both set to use the least congested channel. There is one SSID. What would happen if the non-root bridge did not specify the one and only SSID to be the infrastructure SSID? Will it still connect and handle disassociation/associations correctly?

Everything I'm reading says that you can optionally specify an infrastructure SSID that will force the bridge to use that SSID for the bridge, but it does not say if it's required if you only have one SSID. However, all the examples I see of other peoples bridge configuration is using the command infrastructure-ssid and are also using just one SSID. Which is why I'm wondering if that command is necessary and what it really does when there is only one SSID in range.

cptInsane0
Apr 11, 2007

...and a clown with no head

Martytoof posted:

Truth. I was "Sr. Software Engineer" at the last place I worked and the way they cobbled that title together was the most ridiculous thing seeing as how I had less than a year experience programming at that time, surrounded by people with lesser titles who had been doing it most of their life. For what it's worth I told them I didn't feel right taking the title but they said it was linked to the pay grade so my arguments promptly stopped :)

Bureaucracy! :effort:

That's crazy. Sr. Engineering positions usually have pretty strict requirements. For instance, I still have another year or two before I am qualified for most of those positions. I've been doing this a while.

some kinda jackal
Feb 25, 2003

 
 

cptInsane0 posted:

That's crazy. Sr. Engineering positions usually have pretty strict requirements. For instance, I still have another year or two before I am qualified for most of those positions. I've been doing this a while.

Yeah, I think my company was completely the opposite. They had a database of pay grades and rather than create a new one for me they just shoved me into some preexisting class. Our job titles were basically meaningless. Not bad for a resume though. I don't feel too bad putting it on there since I don't ever plan to go back into Software Engineering so there's no chance I'd ever be hired to work on actual software based on that.

H.R. Paperstacks
May 1, 2006

This is America
My president is black
and my Lambo is blue

Martytoof posted:

Yeah, I think my company was completely the opposite. They had a database of pay grades and rather than create a new one for me they just shoved me into some preexisting class. Our job titles were basically meaningless. Not bad for a resume though. I don't feel too bad putting it on there since I don't ever plan to go back into Software Engineering so there's no chance I'd ever be hired to work on actual software based on that.

Old Job: Sr. Network Engineer
New Job: Engineer Systems 4

I don't touch systems............... I'll never understand titles and why HR has these matrix forms you need to fill into. Exactly why I fall back on the belief my title can be whatever they want it to be as long as my paycheck is the amount I desire.

wolrah
May 8, 2006
what?

routenull0 posted:

Exactly why I fall back on the belief my title can be whatever they want it to be as long as my paycheck is the amount I desire.

Hah, that's some real truth right there.

My official titles through the time I've been at my current job:
Support Technician
Sr. Support Technician (we hired another person)
Network Specialist
LAN/WAN Specialist
Communications Specialist
CIO

My responsibilities really haven't changed significantly, nor has the wide variety of poo poo I end up doing, but every time I got a raise my boss insisted I have a title change.

Somewhere around here I have a small batch of business cards with "Guy who gets poo poo done" in the title field since I jokingly replied to a message asking for my title and the person ordering the cards obviously just hit Forward without reading it.

Tony Montana
Aug 6, 2005

by FactsAreUseless
You went from Support Tech to CIO and you still do support?

Is it a small firm?

wolrah
May 8, 2006
what?

Tony Montana posted:

You went from Support Tech to CIO and you still do support?

Is it a small firm?

<10 employees, like I said my boss just wants me to have a new title with every raise and I just say "sure" because honestly what do I care?

nex
Jul 23, 2001

øæå¨æøåø
Grimey Drawer
I think "Chief consultant" is one of the more absurd titles I see. Dont know why, but it just seems weird.


In other news we have just deployed NetMapper from OPNet(in addition to SP Sentinel with VNE from the same company).
NetMapper can automagically create physical and logical topology maps of your networks which rules when your network grows over a certain size and has a lot of changes. It loving rules. This tool combined with the policy enforcer and reporting server have already made my workday 100x times less frustrating and effective.

An example: Ever set out to trunk a VLAN from one side of your metro to another only to find out that the VLAN id has been taken and someone forgot to update the documentation? Never again, just spit out a updated VLAN map from NetMapper and have a look before starting(and then bitch at someonefor not documenting their work).

Check out some bitching diagrams here:
http://www.opnet.com/solutions/network_planning_operations/netmapper.html

nex fucked around with this message at 10:27 on May 30, 2009

Sojourner
Jun 6, 2007

Get In
Here's one that's stumped the guys at the office for a bit: about 5 or 6 of our switches, older ones (3524s) are no longer accessible via telnet. If you telnet to them you are prompted by a greeting that says
"FreeBSD 4.10 (STABLE)
Kernel 2.6.27 on an i686
login:"
It isn't an IP mixup and we hit a freebsd machine by mistake, its a corrupt IOS so soon I'm going to go up and xmodem a new image, but for curiosity has anyone seen this before? Google came up with nothing.

jwh
Jun 12, 2002

Sojourner posted:

It isn't an IP mixup and we hit a freebsd machine by mistake, its a corrupt IOS so soon I'm going to go up and xmodem a new image, but for curiosity has anyone seen this before? Google came up with nothing.

You're sure about that?

What is the mac address you're seeing for the switch management IP? It might be worth cross checking that with the OUI database.

H.R. Paperstacks
May 1, 2006

This is America
My president is black
and my Lambo is blue

Sojourner posted:

Here's one that's stumped the guys at the office for a bit: about 5 or 6 of our switches, older ones (3524s) are no longer accessible via telnet. If you telnet to them you are prompted by a greeting that says
"FreeBSD 4.10 (STABLE)
Kernel 2.6.27 on an i686
login:"
It isn't an IP mixup and we hit a freebsd machine by mistake, its a corrupt IOS so soon I'm going to go up and xmodem a new image, but for curiosity has anyone seen this before? Google came up with nothing.

I'm with jwh on this one, that isn't a corrupt IOS. 2.6.27 is a recent kernel release for systems, not something you ever find on an old 3524 switch.

tortilla_chip
Jun 13, 2007

k-partite
This sounds more like someone doing ARP spoofing on your management net.

Sojourner
Jun 6, 2007

Get In
It's not arp spoofing, I said the same thing when word of this first got to me. It doesn't make any sense at all really. The intrusion detection system isn't being set off, I plugged myself into the management net and wiresharked it, and just for fun and the sake of using a pricey toy, used our fluke etherscope to resolve all the mac addresses to an IP on the network and found no duplicates. Tomorrow at lunch I'm going to investigate more while people are at lunch. What will come of it, public humiliation or the most bizzare IOS error of all time, stay tuned to find out!

Tremblay
Oct 8, 2002
More dog whistles than a Petco

Sojourner posted:

Here's one that's stumped the guys at the office for a bit: about 5 or 6 of our switches, older ones (3524s) are no longer accessible via telnet. If you telnet to them you are prompted by a greeting that says
"FreeBSD 4.10 (STABLE)
Kernel 2.6.27 on an i686
login:"
It isn't an IP mixup and we hit a freebsd machine by mistake, its a corrupt IOS so soon I'm going to go up and xmodem a new image, but for curiosity has anyone seen this before? Google came up with nothing.

Thirding. Catalyst switches are NOT i686 arch.

jwh
Jun 12, 2002

You can make this easy, just take a machine that's on that network, ping the IP address of your switch management interface. Try and telnet, and make sure you receive the FreeBSD banner. Then arp -an and look at the mac you have for that IP address.

Plug the first 3 bytes into the MAC OUI database lookup: http://standards.ieee.org/regauth/oui/index.shtml

Herv
Mar 24, 2005

Soiled Meat

tortilla_chip posted:

This sounds more like someone doing ARP spoofing on your management net.

For most places, I can use static ARP entries on the important stuff so this doesn't happen. Worst case, when a gateway has to be replaced due to a failure, you just slap the old mac on the new gateway interface.

My :10bux: is on just a dupe IP assignment, not poison arp. Comparing the mac addresses (console cable on the switch I guess) to what's in your arp table sounds like where the trail will start.

I am sure we have all been fooled by less at one time or another. Public Humiliation it is! :)

Herv
Mar 24, 2005

Soiled Meat

jwh posted:

Plug the first 3 bytes into the MAC OUI database lookup: http://standards.ieee.org/regauth/oui/index.shtml

Another :10bux: on the card is a Broadcomm, Intel, or maybe 3com. :)

Oh and 'the trail will start' part meant checking where that mac address is connected to your network/switchport.

Adbot
ADBOT LOVES YOU

Tremblay
Oct 8, 2002
More dog whistles than a Petco

Herv posted:

Another :10bux: on the card is a Broadcomm, Intel, or maybe 3com. :)

Oh and 'the trail will start' part meant checking where that mac address is connected to your network/switchport.

MAC Hunt. Its like Duck Hunt, but less fun. And there isn't an annoying dog to shoot at.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply