The Cisco Short Questions Thread v12.4.9(WTF)

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Cisco Short Questions Thread v12.4.9(WTF)

«‹›431 »

inignot: Sep 1, 2003; WWBCD?

When using etherchannel I've always had only the etherchannel config on the physical ports and left the trunking only on the port channel virtual interface.

# ? Aug 30, 2008 02:25

Adbot: ADBOT LOVES YOU

# ? May 22, 2024 07:52

atticus: Nov 7, 2002; this is how u post~

jbusbysack posted:

As a matter of habit I never let DTP have a crack at anything, as there's no need to negotiate ever between what I would assume is the core switch and a top-of-rack distribution switch. What happens when you change it to 'channel-group 5 mode on' ?

I realize that moving physical ports works for you but I'm curious as to the result.

DTP != aggregation protocols used for Etherchannel.

H110Hawk,

I suggest you try "show etherchannel 5 summary" as you bring up the interfaces so you can monitor the port-states for the ports in Po5 from an etherchannel perspective.

I'd also suggest running "debug etherchannel events" on the 6500 and the relevant equivalent command on the 4948. Everyone likes to freak out about debug commands, but certain ones aren't as bad as others and are completely invaluable troubleshooting tools.

# ? Aug 30, 2008 02:30

ObamaisaTerrist: Jul 26, 2008; The truth is out there.

Can someone fill me in on some basic terminology? Google isn't helping.

Aggregate switch - What do people mean when referring to this?
Core switch - assuming main switch(s) at NOC
Edge switch - assuming last switch before handing off to DMARC

Thanks

# ? Aug 30, 2008 14:46

H.R. Paperstacks: May 1, 2006; This is America
My president is black
and my Lambo is blue

inignot posted:

When using etherchannel I've always had only the etherchannel config on the physical ports and left the trunking only on the port channel virtual interface.

This is the only way I have ever done it.

# ? Aug 30, 2008 16:07

ate shit on live tv: Feb 15, 2004; by Azathoth

ObamaisaTerrist posted:

Can someone fill me in on some basic terminology? Google isn't helping.

Aggregate switch - What do people mean when referring to this?
Core switch - assuming main switch(s) at NOC
Edge switch - assuming last switch before handing off to DMARC

Thanks

Edge switches are smaller less powerful switches that the users actually plug directly into (2900's, etc).

Aggregation are better switches that can do MPLS and usually have fiber uplinks to the core (3560, 3750, 3400 Metro, 4948's).

Core are the big switches like 6500s and GSRs (sometimes a 4948). They usually hand off to the main routers 7600, and CRSs.

Aggregation isn't always needed on an individual business level. On the Service Provider level the Core is typically made up of routers and the aggregation is 6500's full of 6748 SFPs, with some ES20's thrown in for full line rate handoff to the core.

But like everything else in networking this is just a guideline, and there is no hard and fast definition for separating core from edge from aggregation.

# ? Aug 30, 2008 17:12

atticus: Nov 7, 2002; this is how u post~

http://en.wikipedia.org/wiki/Hierarchical_internetworking_model

Feel free to use the terms "access" and "edge" interchangeably, although lately I've seen "edge" be the more popular term.

# ? Aug 30, 2008 21:28

jbusbysack: Sep 6, 2002; i heart syd

atticus posted:

DTP != aggregation protocols used for Etherchannel.

H110Hawk,

I suggest you try "show etherchannel 5 summary" as you bring up the interfaces so you can monitor the port-states for the ports in Po5 from an etherchannel perspective.

I'd also suggest running "debug etherchannel events" on the 6500 and the relevant equivalent command on the 4948. Everyone likes to freak out about debug commands, but certain ones aren't as bad as others and are completely invaluable troubleshooting tools.

Wow, I dunno why I called that DTP, but my question remains. Why allow it to negotiate instead of forcing the port bundling?

# ? Aug 31, 2008 04:36

ior: Nov 21, 2003; What's a fuckass?

jbusbysack posted:

Wow, I dunno why I called that DTP, but my question remains. Why allow it to negotiate instead of forcing the port bundling?

Because when someone makes a mistake while patching and patches a users laptop into the channel it will effectively blackhole 50% of the traffic going through it. Happy times were had troubleshooting why someone could access website x but not y while someone else could access y but not x.

# ? Aug 31, 2008 10:30

ObamaisaTerrist: Jul 26, 2008; The truth is out there.

Thanks for the info.

In continuing my questions, how do you all utilize the sh mac-address-table command? I understand what it does, I just haven't (yet) run into a day to day scenario when I would need to use it. Can you show me how it helps you? Thanks

# ? Aug 31, 2008 16:37

jbusbysack: Sep 6, 2002; i heart syd

ObamaisaTerrist posted:

Thanks for the info.

In continuing my questions, how do you all utilize the sh mac-address-table command? I understand what it does, I just haven't (yet) run into a day to day scenario when I would need to use it. Can you show me how it helps you? Thanks

Tracking down which access (edge) switch a user and their associated IP address is plugged into from the core. Basically comparing arp / mac-add table and hopping around out various uplinks until you find the access switch/port that the user is in.

This obviously doesn't work from a routed perspective but on local segments it is very helpful.

# ? Aug 31, 2008 18:28

FatCow: Apr 22, 2002; I MAP THE FUCK OUT OF PEOPLE

I've also used it when I have a server that was provisioned wrong. With 'sh mac-address-table' and '(config)# arp' you can force the MAC into the ARP tables to recover the box.

# ? Aug 31, 2008 20:30

ate shit on live tv: Feb 15, 2004; by Azathoth

If there is a duplicate IP on the network and you can't track down what device it is, you can use 'sh mac-address-table' plus 'show arp' and figure out the physical port the duplicate IP is on. Then you can manually trace the wire to the device.

I had to do this a few times when I was a Lab COOP. 4 dynamic Class C networks, with about 160 devices each made the likely hood of duplicate addresses common.

# ? Sep 1, 2008 02:59

jwh: Jun 12, 2002

ObamaisaTerrist posted:

In continuing my questions, how do you all utilize the sh mac-address-table command? I understand what it does, I just haven't (yet) run into a day to day scenario when I would need to use it.

I use this frequently, typically in response to "can you help, we don't know where we plugged this server in. It's name is <server name>". I then ping server name from a machine that's resolving towards internal DNS servers, grab the IP, and ping from a switch. Then I'll show arp | i <IP>, and copy the mac address seen. Then I'll "sh mac-address-table | i <mac address>", and repeat until I'm on the servicing switch for that particular machine.

I do this probably a half dozen times a week. I suppose we could circumvent the entire process by requiring the server teams to keep track of what they plug into where, but you know how it goes.

Did anyone ever come upon a CDP agent for windows machines that isn't some weird RBN malware deathtrap?

# ? Sep 1, 2008 06:05

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

jwh posted:

Did anyone ever come upon a CDP agent for windows machines that isn't some weird RBN malware deathtrap?

Haven't used this, but it seems to have a bunch of good shareware reviews, and it has a Windows service as well as a client to view CDP info (so the server admins can just pull that up to see where they're plugged in at) : http://www.tallsoft.com/cdpciscoclient.htm

# ? Sep 1, 2008 13:46

H110Hawk: Dec 28, 2006

atticus posted:

I suggest you try "show etherchannel 5 summary" as you bring up the interfaces so you can monitor the port-states for the ports in Po5 from an etherchannel perspective.

I'd also suggest running "debug etherchannel events" on the 6500 and the relevant equivalent command on the 4948. Everyone likes to freak out about debug commands, but certain ones aren't as bad as others and are completely invaluable troubleshooting tools.

I'll try this out with a cold 4948 and the bum port next week sometime.

We use desirable for the other quoted reasons. We used to use `on`, but it's just too annoying having some douchebag with a laptop getting 500mbit of traffic destined for somewhere else.

And we use the exact setup mentioned for core/dist(aggregation)/edge. 6509 with etherchanneled 4948's hooked up to it, and single-gig 2960's which have the servers themselves plugged in to them. Yes, 100mbit to the servers. We tend to use 3560G's for the filer edge, though.

# ? Sep 1, 2008 17:16

tortilla_chip: Jun 13, 2007; k-partite

You can also use a L2 traceroute (there are some caveats)

http://www.cciecandidate.com/?p=513

# ? Sep 1, 2008 18:49

Chuu: Sep 11, 2004; Grimey Drawer

I have a rj-45 -> db9 management cable. If I use a db9->usb adaptor (example : link) can I use this on a computer without a serial port?

I'm asking because it looks like these are "dumb" adapters and not truly emulating a serial port. If the answer is no, any suggestions on how to search eBay for this in a way that's going to filter all of those out?

# ? Sep 1, 2008 20:47

n0tqu1tesane: May 7, 2003; She was rubbing her ass all over my hands. They don't just do that for everyone.; Grimey Drawer

Chuu posted:

I have a rj-45 -> db9 management cable. If I use a db9->usb adaptor (example : link) can I use this on a computer without a serial port?

I'm asking because it looks like these are "dumb" adapters and not truly emulating a serial port. If the answer is no, any suggestions on how to search eBay for this in a way that's going to filter all of those out?

I've had some coworkers who have had to do this because their serial port died, and yes it will work.

# ? Sep 1, 2008 20:53

H110Hawk: Dec 28, 2006

Chuu posted:

I have a rj-45 -> db9 management cable. If I use a db9->usb adaptor (example : link) can I use this on a computer without a serial port?

You can't really have a "dumb" adapter. A USB and a RS-232 chip are both pennies per unit. You are looking at the finest in slave labor assembled/stolen/both USB adapter. I would suggest getting a keyspan.

# ? Sep 1, 2008 21:59

Alowishus: Jan 8, 2002; My name is Mud

I have an 1841 router that shipped with version 12.3(8)T IOS software and 12.3(8r)T8 ROM (this is what Cisco calls "ROMMON" right?)

If I want to update my software to the latest 12.4(20)T, should I also update the ROMMON? Looking through my support portal, version 12.4(13r)T5 is available for download. Is this even possible via software flash? I can't seem to find any sort of instructions about how to do it. Should I care?

# ? Sep 3, 2008 02:53

jwh: Jun 12, 2002

You shouldn't need to update your ROMMON to boot a new IOS on the 1841.

# ? Sep 3, 2008 03:02

Alowishus: Jan 8, 2002; My name is Mud

No, you're right, I tested it and 12.4 boots fine. Was just wondering if there was any need for or benefit from also updating the ROMMON.

Edit: also thanks for your earlier help about my XO routing questions... I wound up taking the 802.1Q approach and so far so good.

# ? Sep 3, 2008 03:22

jwh: Jun 12, 2002

Alowishus posted:

No, you're right, I tested it and 12.4 boots fine. Was just wondering if there was any need for or benefit from also updating the ROMMON.

I've seen weird stuff on newer 871s where they won't boot an older IOS, and I suspect it has to do with the ROMMON version, but I never figured it out completely. I can tell you that I've never had to update a ROMMON ever, except maybe once, and my memory of that isn't very good.

quote:

Edit: also thanks for your earlier help about my XO routing questions... I wound up taking the 802.1Q approach and so far so good.

Oh cool, well that's good to hear. I'm glad it worked out!

# ? Sep 3, 2008 05:31

Chuu: Sep 11, 2004; Grimey Drawer

Two part question here. First part is about some basic config stuff. I bought some 2500 series routers to study for the CCNA. I have two 2501's and a 2503. The 2501's have 2 serial ports and an AUI. The 2503 has the same, plus an ISDN BRI port.

The 2503 does something the 2501's don't, which I guess is related to the BRI port. On bootup, it displays:

code:

Copyright (c) 1986-1996 by cisco Systems
2500 processor with 14336 Kbytes of main memory

%SYS-4-CONFIG_NEWER: Configurations from version 12.2 may not be correctly understood.

--->output remove---<

ip kerberos source-interface any
   ^

% Invalid input detected at '^' marker.
--->lots of output removed<---

It's a 16/16 router, so 2 megs of memory are missing, and that text message means that probably some module doing boot is eating 2 megs of memory and running it's own configuration. I am assuming that configuration is in the private-config file in the nvram. My question basically is, what happens if I nuke the nvram? Are things going to break horribly? Also, how do I look at that private-config? I can see it using the dir commands, but I can't seem to find a 'show' command to let me view it.

Part 2

One of the 2501's had some really ancient IOS version that didn't like many commands in the Odom book, so I flashed it to the newest version supported. Everything went fine.

Then I try to flash this 2503. It's just not working though. The tftpd log shows me that it connects fine, and I can upload fine. When I try to download though, I get this:

code:

Connection received from 192.168.0.1 on port 53101 [02/09 23:51:25.281]
Read request for file <c2500-is-l.123-20.bin>. Mode octet [02/09 23:51:25.281]
Using local port 2532 [02/09 23:51:25.281]
Peer returns ERROR <Buffer overflow> -> aborting transfer [02/09 23:51:25.343]
Warning : received duplicated request [02/09 23:51:25.390]
Warning : Unaccepted request received from 192.168.0.1 [02/09 23:51:25.390]
Connection received from 192.168.0.1 on port 57271 [02/09 23:51:29.390]
Read request for file <c2500-is-l.123-20.bin>. Mode octet [02/09 23:51:29.390]

I have the same problem when trying to connect to the tftp server directly from the bootstrap loader. Is there some way around this to get a newer IOS image on this router?

# ? Sep 3, 2008 06:46

H110Hawk: Dec 28, 2006

Chuu posted:

I have the same problem when trying to connect to the tftp server directly from the bootstrap loader. Is there some way around this to get a newer IOS image on this router?

You kids and your trivial file transfer protocol witchcraft. Have you tried Xmodem, or similar?

# ? Sep 3, 2008 07:08

Chuu: Sep 11, 2004; Grimey Drawer

H110Hawk posted:

You kids and your trivial file transfer protocol witchcraft. Have you tried Xmodem, or similar?

Nope, will try it.

# ? Sep 3, 2008 07:33

H.R. Paperstacks: May 1, 2006; This is America
My president is black
and my Lambo is blue

jwh posted:

I've seen weird stuff on newer 871s where they won't boot an older IOS, and I suspect it has to do with the ROMMON version, but I never figured it out completely. I can tell you that I've never had to update a ROMMON ever, except maybe once, and my memory of that isn't very good.

Oh cool, well that's good to hear. I'm glad it worked out!

There are occasions on older 2600 series routers and even some newer stuff that ROMMON/Bootstrap has to be upgraded in order to support larger memory modules. I remember having to do this specifically on a 2620XM.

# ? Sep 3, 2008 12:14

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

Chuu posted:

Nope, will try it.

xmodem is horribly slow, if you have a network connection available, and a unix box, try using Shrubbery networks RCPd, it's intended as a drop-in replacement for tftp since tftp sucks: http://www.shrubbery.net/rcpd/

# ? Sep 3, 2008 12:39

inignot: Sep 1, 2003; WWBCD?

Of course if you are reduced to xmodeming code onto a router, remember to turn the serial port speed on the router & your workstation up to the maximum supported value. It's still going to be horrible though.

# ? Sep 3, 2008 16:09

atticus: Nov 7, 2002; this is how u post~

H110Hawk posted:

You kids and your trivial file transfer protocol witchcraft. Have you tried Xmodem, or similar?

wtf

# ? Sep 3, 2008 16:22

Chuu: Sep 11, 2004; Grimey Drawer

Girdle Wax posted:

xmodem is horribly slow, if you have a network connection available, and a unix box, try using Shrubbery networks RCPd, it's intended as a drop-in replacement for tftp since tftp sucks: http://www.shrubbery.net/rcpd/

RCP didn't come into IOS until 12.3. These routers are loaded with 12.2 and 12.1, I thought tftp or direct serial was the only option, didn't realize you could use xmodem from the bootstrap loader.

# ? Sep 3, 2008 21:57

H110Hawk: Dec 28, 2006

Chuu posted:

RCP didn't come into IOS until 12.3. These routers are loaded with 12.2 and 12.1, I thought tftp or direct serial was the only option, didn't realize you could use xmodem from the bootstrap loader.

Oh, shoot. I meant to say the slowest option. Use serial set to lowest baud rate.

# ? Sep 4, 2008 02:33

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

Chuu posted:

RCP didn't come into IOS until 12.3. These routers are loaded with 12.2 and 12.1, I thought tftp or direct serial was the only option, didn't realize you could use xmodem from the bootstrap loader.

code:

#copy ?
  /erase          Erase destination file system.
  /noverify       Don't verify image signature before reload.
  /verify         Verify image signature before reload.
  bootflash:      Copy from bootflash: file system
  cns:            Copy from cns: file system
  disk0:          Copy from disk0: file system
  disk1:          Copy from disk1: file system
  flash:          Copy from flash: file system
  ftp:            Copy from ftp: file system
  null:           Copy from null: file system
  nvram:          Copy from nvram: file system
  rcp:            Copy from rcp: file system
  running-config  Copy from current system configuration
  scp:            Copy from scp: file system
  slot0:          Copy from slot0: file system
  slot1:          Copy from slot1: file system
  startup-config  Copy from startup configuration
  system:         Copy from system: file system
  tar:            Copy from tar: file system
  tftp:           Copy from tftp: file system

#show ver
Cisco Internetwork Operating System Software 
IOS (tm) GS Software (GSR-K4P-M), Version 12.0(28)S, EARLY DEPLOYMENT RELEASE SOFTWARE (fc1)
Technical Support: [url]http://www.cisco.com/techsupport[/url]
Copyright (c) 1986-2004 by cisco Systems, Inc.
Compiled Mon 03-May-04 16:46 by nmasa
Image text-base: 0x50010E50, data-base: 0x535CC000

Granted it's S train, but still, RCP support has been around for awhile.

-edit-
12.0 ios fundamentals mentions rcp: http://www.cisco.com/en/US/docs/ios/12_0/configfun/configuration/guide/fccfgfil.html
-/edit-

ragzilla fucked around with this message at 02:42 on Sep 4, 2008

# ? Sep 4, 2008 02:40

jwh: Jun 12, 2002

Not exactly Cisco related per se, but I finally got my hands on a Lantronix SLC eval unit- it's a console server that runs a Linux kernel, and has some pretty fancy bells and whistles. For what it's worth, it's much nicer than the Avocent console servers we've used here in the past.

What I need, however, is a bunch of 568B to rollover couplers- has anybody found anything like this? I'd rather spend money than cut cable ends.

# ? Sep 4, 2008 18:20

Eyecannon: Mar 13, 2003; you are what you excrete

Cisco n00b here, trying to fix our VPN at work. It's a 1811W, and I'm dealing with it through SDM.

It just needs to be a simple Easy VPN Server. I want the clients to be able to access the entire internal network (10.0.0.0/24), but right now, they can only ping the internal address of the Cisco itself (10.0.0.5).

Where should I be looking to allow it to see the entire subnet?

Eyecannon fucked around with this message at 19:11 on Sep 4, 2008

# ? Sep 4, 2008 19:08

jwh: Jun 12, 2002

Eyecannon posted:

Where should I be looking to allow it to see the entire subnet?

Are the other machines on your network routing back to the 1811 correctly? Is there more than one gateway on this network?

# ? Sep 4, 2008 19:35

Eyecannon: Mar 13, 2003; you are what you excrete

jwh posted:

Are the other machines on your network routing back to the 1811 correctly? Is there more than one gateway on this network?

Yes, from any machine internally I can access it.

Yes I guess we do have two gateways. The Cisco's external network is x.x.x.217 (this is FE0), but there is also a gateway/firewall at 10.0.0.1 that our internal traffic goes out, and it talks to an external x.x.x.218 address. So all the internal workstations here have 10.0.0.1 as their gateway. I don't want the Cisco to be the gateway for our internal stuff, just that if someone VPN's in on x.x.x.217, they can access the entire 10.0.0.0/24 subnet.

# ? Sep 4, 2008 20:04

jwh: Jun 12, 2002

When someone VPNs into the Cisco, what IP addresses are they being assigned?

# ? Sep 4, 2008 22:13

Eyecannon: Mar 13, 2003; you are what you excrete

jwh posted:

When someone VPNs into the Cisco, what IP addresses are they being assigned?

They get something in 192.168.1.0/24.

# ? Sep 4, 2008 22:17

Adbot: ADBOT LOVES YOU

# ? May 22, 2024 07:52

jwh: Jun 12, 2002

Okay, so you'll need to add a route to your "real" gateway to route 192.168.1.0/24 towards the 10.0.0.5 IP address of your 1800.

# ? Sep 4, 2008 22:37

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Cisco Short Questions Thread v12.4.9(WTF)

«‹›431 »