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


zalmoxes posted:

I'm probably getting ahead of myself, but I'm just studying for the CCNA and wanted to know what happens to UDP traffic in a setup with load-ballancing? UDP doesn't offer any sort of error-recovery or reordering of packets, so if the traffic gets load ballanced over unequal links, how come it doesn't become scrambled on the other end?
Usually the application will use a sequence number and buffer/reorder (see jitter buffers common in RTP applications), or the data exchanged is small enough that it only uses a single datagram.

Adbot
ADBOT LOVES YOU

Jelmylicious
Dec 6, 2007
Buy Dr. Quack's miracle juice! Now with patented H-twenty!
This is also why voice and video use RTP on top of UDP: http://en.wikipedia.org/wiki/Real-time_Transport_Protocol

e:fb

Ninja Rope
Oct 22, 2005

Wee.

zalmoxes posted:

I'm probably getting ahead of myself, but I'm just studying for the CCNA and wanted to know what happens to UDP traffic in a setup with load-ballancing? UDP doesn't offer any sort of error-recovery or reordering of packets, so if the traffic gets load ballanced over unequal links, how come it doesn't become scrambled on the other end?

Regarding LACP, I think it's the Ethernet standard that says devices should avoid re-ordering packets whenever possible. That's part of the reason LACP doesn't scale exactly with the number of links added. LACP uses a hash algorithm to determine which link an outgoing packet will be sent over, and that hash algorithm is designed to keep a single flow on the same link, even if that results in unused capacity. There are multiple algorithm choices on some platforms but they're all based around how you determine what a flow is.

In general, though, you can set up some kind of hosed up load balancing scenario that causes packets to be delivered out of order, and in the case of UDP (or any protocol that doesn't build in retransmits and reordering) it is possible for lost data or delayed data to cause problems. So don't do that!

lilcasino
Jun 4, 2001
Anyone recommend a Cisco router for our new fiber connection? We're going from bonded T's to a fiber ethernet hand off. The only "weird" part the ISP is giving me one subnet for the router and another subnet for our internal IP's

falz
Jan 29, 2005

01100110 01100001 01101100 01111010

lilcasino posted:

Anyone recommend a Cisco router for our new fiber connection? We're going from bonded T's to a fiber ethernet hand off. The only "weird" part the ISP is giving me one subnet for the router and another subnet for our internal IP's
What speed and type of media is the handoff? What features do you need (Nat?) Also using a link net to route your block is nice since it will give you more usable IPs.

lilcasino
Jun 4, 2001

falz posted:

What speed and type of media is the handoff? What features do you need (Nat?) Also using a link net to route your block is nice since it will give you more usable IPs.

Fairpoint has us a 10m right now but it is scalable to 100 if we need it. It is a straight RJ45 ethernet hand off. We average around 2500 connections at one time. I don't need the Cisco to do NAT my firewall would do that.

ate shit on live tv
Feb 15, 2004

by Azathoth
Depending on the features you need (Wireless etc) you can get away with a simple 888 or 1800.

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


lilcasino posted:

Fairpoint has us a 10m right now but it is scalable to 100 if we need it. It is a straight RJ45 ethernet hand off. We average around 2500 connections at one time. I don't need the Cisco to do NAT my firewall would do that.

Why do you want/need a router? If they're going to hand off Ethernet why not take it direct into your firewall, unless you have a need for something in a DMZ outside the firewall?

lilcasino
Jun 4, 2001

ragzilla posted:

Why do you want/need a router? If they're going to hand off Ethernet why not take it direct into your firewall, unless you have a need for something in a DMZ outside the firewall?

Fairpoint has given us a /52 for our router then a /40. I believe the only way I can route the /40 with our firewall is to use ARP and create static routes for each address. Wouldn't it be easier to configure and maintain to have a router on the /52 and then pass the /40 subnet? Or am I over or not thinking.

Ginger Beer Belly
Aug 18, 2010



Grimey Drawer

lilcasino posted:

Fairpoint has given us a /52 for our router then a /40. I believe the only way I can route the /40 with our firewall is to use ARP and create static routes for each address. Wouldn't it be easier to configure and maintain to have a router on the /52 and then pass the /40 subnet? Or am I over or not thinking.

Do you mean a 255.255.255.252 mask subnet and then a 255.255.255.240 mask subnet?

You may want to check out http://www.subnet-calculator.com/cidr.php

What is your firewall?

Tasty Wheat
Jul 18, 2012

Ninja Rope posted:

Regarding LACP, I think it's the Ethernet standard that says devices should avoid re-ordering packets whenever possible. That's part of the reason LACP doesn't scale exactly with the number of links added. LACP uses a hash algorithm to determine which link an outgoing packet will be sent over, and that hash algorithm is designed to keep a single flow on the same link, even if that results in unused capacity. There are multiple algorithm choices on some platforms but they're all based around how you determine what a flow is.

In general, though, you can set up some kind of hosed up load balancing scenario that causes packets to be delivered out of order, and in the case of UDP (or any protocol that doesn't build in retransmits and reordering) it is possible for lost data or delayed data to cause problems. So don't do that!

At least from the WAN side going in, voice can get really jacked up if you just load balance everything without regard to the type of traffic. TCP works just fine, even when your setup is 2 3.8m VSATs connected by 1km of SM, voice not so much.

falz
Jan 29, 2005

01100110 01100001 01101100 01111010

lilcasino posted:

Fairpoint has given us a /52 for our router then a /40. I believe the only way I can route the /40 with our firewall is to use ARP and create static routes for each address. Wouldn't it be easier to configure and maintain to have a router on the /52 and then pass the /40 subnet? Or am I over or not thinking.

I don't see why a firewall couldn't do a /30 on one ethernet interface and your /29 or whatever inside. What type of firewall is it?

Sir Sidney Poitier
Aug 14, 2006

My favourite actor


lilcasino posted:

Anyone recommend a Cisco router for our new fiber connection? We're going from bonded T's to a fiber ethernet hand off. The only "weird" part the ISP is giving me one subnet for the router and another subnet for our internal IP's

It's probably overkill, but this is the setup we have for our leased line customers who have a fibre for the main connection, then ADSL as backup. We send out a 1941 for it with an HWIC1-ADSL or whatever it is. This then speaks BGP with our PE and advertises the customer network over both, but lower pref on the ADSL.

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


falz posted:

I don't see why a firewall couldn't do a /30 on one ethernet interface and your /29 or whatever inside. What type of firewall is it?

Typically I'd do /30 on the outside (presumably the provider has the /29 routed to the customer side of the /30) and all RFC1918 on the inside, then you can use all 8 IPs of the /29 for translations.

brent78
Jun 23, 2004

I killed your cat, you druggie bitch.
Cisco 2911/K9 (C2900-UNIVERSALK9-M), Version 15.1(4)M4, RELEASE SOFTWARE (fc1)

Shouldn't this be supported?

code:
(config)#ip urlfilter server vendor websense 172.16.4.12 timeout 8 retransmit 6
                        ^
% Invalid input detected at '^' marker.

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


brent78 posted:

Cisco 2911/K9 (C2900-UNIVERSALK9-M), Version 15.1(4)M4, RELEASE SOFTWARE (fc1)

Shouldn't this be supported?

code:
(config)#ip urlfilter server vendor websense 172.16.4.12 timeout 8 retransmit 6
                        ^
% Invalid input detected at '^' marker.

Does urlfilter require security license/image?

ToG
Feb 17, 2007
Rory Gallagher Wannabe
I have a question regarding NAT.

I currently have two networks

10.200.0.0/24 which is for my Cisco/Virtual stuff
192.168.1.0/24 which is my home network.

I currently have internet access via the home network. I have a Cisco 2621XM sitting on both networks and am using NAT overload to get internet access for the cisco/virtual stuff.

This works perfectly at the minute, but I'd like to forward a port on the home network to a virtual machine and vice versa. The home router can ONLY do port forwarding.

Whenever I create a static PAT for the port I want to the virtual machine, The virtual machine loses it's internet connection, is there a way to forward this port as well as having NAT overload in place?

edit: I'm using the command:

ip nat outside source static tcp 192.168.1.1 PORT 10.200.0.110 PORT extendable

edit2: Relevant config in place
code:
!
interface FastEthernet0/0
 description ## INTERNET ##
 ip address 192.168.1.1 255.255.255.0
 ip nat outside
 ip virtual-reassembly
 speed 100
 full-duplex
!
interface FastEthernet0/1
 no ip address
 ip nat inside
 ip virtual-reassembly
 speed 100
 full-duplex
!
interface FastEthernet0/1.30
 encapsulation dot1Q 30
 ip address 10.200.0.97 255.255.255.224
 ip nat inside
 ip virtual-reassembly

ip route 0.0.0.0 0.0.0.0 FastEthernet0/0 192.168.1.254 name HomeRouter
!
ip nat inside source list 1 interface FastEthernet0/0 overload
!
access-list 1 permit 10.200.0.96 0.0.0.31
!

ToG fucked around with this message at 08:53 on Jul 24, 2012

Sepist
Dec 26, 2005

FUCK BITCHES, ROUTE PACKETS

Gravy Boat 2k
Change

ip nat outside source static tcp 192.168.1.1 PORT 10.200.0.110 PORT extendable

to

ip nat inside source static tcp 192.168.1.1 PORT 10.200.0.110 PORT extendable

ToG
Feb 17, 2007
Rory Gallagher Wannabe

Sepist posted:

Change

ip nat outside source static tcp 192.168.1.1 PORT 10.200.0.110 PORT extendable

to

ip nat inside source static tcp 192.168.1.1 PORT 10.200.0.110 PORT extendable

Sorry, I had meant that I just transposed it wrong. Doing that cuts off the internet connection to the virtual machine too

Sepist
Dec 26, 2005

FUCK BITCHES, ROUTE PACKETS

Gravy Boat 2k
Ah, noticed the inside IP should be the first with the interface being the second.

ip nat inside source static tcp 10.200.0.110 PORT interface FastEthernet0/0 PORT

falz
Jan 29, 2005

01100110 01100001 01101100 01111010

ragzilla posted:

Does urlfilter require security license/image?
Probably. http://www.cisco.com/en/US/prod/collateral/modules/ps10598/ordering_guide_c07_557736_ps10536_Products_Data_Sheet.html#wp9000806

ToG
Feb 17, 2007
Rory Gallagher Wannabe

Sepist posted:

Ah, noticed the inside IP should be the first with the interface being the second.

ip nat inside source static tcp 10.200.0.110 PORT interface FastEthernet0/0 PORT

drat, How'd I miss that, I configured the nat overload like a day ago. Thanks :D . Will test it when I get a chance.


edit: Tested and it performs perfectly. Thanks again.

ToG fucked around with this message at 15:23 on Jul 25, 2012

Sir Sidney Poitier
Aug 14, 2006

My favourite actor


If you try and advertise a prefix to a peer, but the peer doesn't allow it due to a prefix filter, will "show ip bgp neighbor X.X.X.X advertised-routes" show the prefix? Or does it only show those which are accepted?

Ginger Beer Belly
Aug 18, 2010



Grimey Drawer

Anjow posted:

If you try and advertise a prefix to a peer, but the peer doesn't allow it due to a prefix filter, will "show ip bgp neighbor X.X.X.X advertised-routes" show the prefix? Or does it only show those which are accepted?

It will show up.

If you have access to the peer, you can look at "sh ip bgp neigh X.X.X.X received-routes" and compare it to "sh ip bgp neigh X.X.X.X routes" to see what was advertised vs what was accepted and installed.

zalmoxes
Sep 30, 2009

:eurovision:
I use a NM-16A module on a router to console into other routers and switches
code:
 
RC3745#show run | inc ip host
ip host sw1 2129 192.168.10.10
ip host sw2 2130 192.168.10.10
ip host sw3 2131 192.168.10.10
I'm having several issues with this. First of all, if I reboot everything and want to telnet into a switch from the router, it doesn't work. until I clear the line and try again. Then I get a lot of garbled output, I'm assuming the console cable is transmitting on all the lines at the same time? I'm not sure how to fix this.

Here is how it looks.

code:
RC3745#sw2
Translating "sw2"
Trying sw2 (192.168.10.10, 2130)...
% Connection refused by remote host

RC3745#clear line 130
[confirm]
 [OK]  
RC3745#sw2
Translating "sw2"
Trying sw2 (192.168.10.10, 2130)... Open        $own command or computer name, 
or unable to find computer address

sw2>$known command or computer name, or unable t$known command or computer name, 
^ or unable to find computer address   

% Invalid input detected at '^' marker.

sw2>   

zalmoxes fucked around with this message at 20:16 on Jul 26, 2012

Tasty Wheat
Jul 18, 2012

zalmoxes posted:

I use a NM-16A module on a router to console into other routers and switches
code:
 
RC3745#show run | inc ip host
ip host sw1 2129 192.168.10.10
ip host sw2 2130 192.168.10.10
ip host sw3 2131 192.168.10.10
I'm having several issues with this. First of all, if I reboot everything and want to telnet into a switch from the router, it doesn't work. until I clear the line and try again. Then I get a lot of garbled output, I'm assuming the console cable is transmitting on all the lines at the same time? I'm not sure how to fix this.

Here is how it looks.

code:
RC3745#sw2
Translating "sw2"
Trying sw2 (192.168.10.10, 2130)...
% Connection refused by remote host

RC3745#clear line 130
[confirm]
 [OK]  
RC3745#sw2
Translating "sw2"
Trying sw2 (192.168.10.10, 2130)... Open        $own command or computer name, 
or unable to find computer address has bee

sw2>$known command or computer name, or unable t$known command or computer name, 
^ or unable to find computer address   

% Invalid input detected at '^' marker.

sw2>   

I have both a AS2511-RJ and NM-32A, my personal experience has been the same, being that I am lazy, I just copy and paste out of a text file all the clear line commands.

Sepist
Dec 26, 2005

FUCK BITCHES, ROUTE PACKETS

Gravy Boat 2k

Ginger Beer Belly posted:

It will show up.

If you have access to the peer, you can look at "sh ip bgp neigh X.X.X.X received-routes" and compare it to "sh ip bgp neigh X.X.X.X routes" to see what was advertised vs what was accepted and installed.

Caveat emptor, received-routes only works if soft-configuration is enabled for the peer.

Ridge
Feb 22, 2004
it's a tarp

lol internet. posted:

Quick question, for Cisco ASA. Does anyone use the CLI to configure\manage access rules? Or is everyone using the ASDM?

I use ASDM sometimes for debugging, but that's it. Once I used ASDM to create a VPN and some access rules. I was not impressed with the mess it dumped in the config.

zalmoxes
Sep 30, 2009

:eurovision:

quote:

All Catalyst multilayer switches support the following types of layer 3 interfaces:

Routed port – a pure layer 3 port similar to that on a router
Switch virtual interface (SVI) – virtual routed VLAN interface for inter-VLAN routing
Bridge virtual interface (BVI) – a layer 3 bridging interface

I understand what an SVI is, but what is the difference between a 'routed port' and a 'BVI' on a Catalyst L3 switch.

When I want to communicate with a router on L3 I type in
int fa0/0
no switchport
ip address xxx.xxx.xxx.xxx xxx.xxx.xxx.xxx

and I have a routed port.

When I need a Vlan Interface, I type in
int Vlan20
ip address xxx.xxx.xxx.xxx xxx.xxx.xxx.xxx

and I create a SVI

What is a BVI?

BurgerQuest
Mar 17, 2009

by Jeffrey of YOSPOS

zalmoxes posted:

I understand what an SVI is, but what is the difference between a 'routed port' and a 'BVI' on a Catalyst L3 switch.

When I want to communicate with a router on L3 I type in
int fa0/0
no switchport
ip address xxx.xxx.xxx.xxx xxx.xxx.xxx.xxx

and I have a routed port.

When I need a Vlan Interface, I type in
int Vlan20
ip address xxx.xxx.xxx.xxx xxx.xxx.xxx.xxx

and I create a SVI

What is a BVI?

I've never used it but this seemed to explain it ok to me:

http://www.cisco.com/en/US/tech/tk389/tk815/technologies_tech_note09186a0080094663.shtml

Basically it lets you provide a virtual interface that exists within a bridge-group of 2 or more routed ports.



It seems BVI is not really used anymore in favour of SVI, as I don't think there's any advantage in using routed ports in this fashion over VLAN/SVI? It essentially turns a router working like a switch, except split into bridge-groups instead of VLANs.

BurgerQuest fucked around with this message at 03:33 on Jul 31, 2012

FatCow
Apr 22, 2002
I MAP THE FUCK OUT OF PEOPLE
It's used to make a router appear somewhat switch like. You don't get SVIs in router land.

CrazyLittle
Sep 11, 2001





Clapping Larry
SVI's don't do traffic shaping, from what I've encountered.

ate shit on live tv
Feb 15, 2004

by Azathoth
SVIs are also virtual ports, whereas a Routed port is physical, meaning it his the L3 ASICs.

EIGRP recovers faster over a pair of Routed Ports rather than an SVI. Also QoS/Shaping.

http://packetlife.net/blog/2011/jan/24/convergence-delays-svi-vs-routed-interface/

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


Powercrazy posted:

SVIs are also virtual ports, whereas a Routed port is physical, meaning it his the L3 ASICs.

EIGRP recovers faster over a pair of Routed Ports rather than an SVI. Also QoS/Shaping.

http://packetlife.net/blog/2011/jan/24/convergence-delays-svi-vs-routed-interface/
internally there's no difference between a routed port and terminating on SVI, the routed port will assign an internal VLAN and it just hides the SVI. The one speed up you do get is that it skips checking if there are any other ports in the VLAN up before signaling the SVI to go down if you use a routed interface since it knows there's only 1 port in the VLAN.

Gap In The Tooth
Aug 16, 2004
Never done voice before, but now have two 7940 IP phones.

My understanding is if I put the voice version of the IOS on my 2620 router, and then get a copy of UCM Express and put it on the router I should be able to get the phones to talk to each other given the correct settings.

Is this more or less what I need to get or are there elements I'm missing?

Partycat
Oct 25, 2004

As far as I know with the appropriate software, CME, the phones can talk to each other once it is configured, as it will do everything it needs to do internally.

If your phones are SCCP anyways.

CrazyLittle
Sep 11, 2001





Clapping Larry
augh, why use skinny in 2012?

ate shit on live tv
Feb 15, 2004

by Azathoth

CrazyLittle posted:

augh, why use skinny in 2012?

Calm down, many people don't know any better, but of course if that is the case they shouldn't be deploying phones in the enterprise environment :eng99:

Partycat
Oct 25, 2004

Well, we already drank the Cisco Kool-Aid, so I have 8000+ sets running SCCP.

Converting them to SIP themselves seems pointless at this time.

Adbot
ADBOT LOVES YOU

adorai
Nov 2, 2002

10/27/04 Never forget
Grimey Drawer
What should I be deploying them as instead of SCCP?

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