Drag to reposition cover

Brocade ICX Series (cheap & powerful 10gbE/40gbE switching)

Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.

dashpuppy

Member
Dec 16, 2018
48
21
8
Guys, any help here.. would be good :p

I got my 6450-24p switch. Factory defaulted it and setup the username and password following instructions on the internet. BUT when i go to the web-page of the swtich all i see is the switch picture.

I have tried different browsers rebooted witch etc etc.

Trying to enable the web-gui so i can look at the switch then update it and start configuring it. IMO the cli is really close to Cisco.. Glad i kept my cisco consol cables & usb to serial Adapters.

ICX6450-24P Switch>show ver
Copyright (c) 1996-2014 Brocade Communications Systems, Inc. All rights reserv ed.
UNIT 1: compiled on Dec 17 2014 at 17:45:46 labeled as ICX64S08010g
(8104988 bytes) from Primary ICX64S08010g.bin
SW: Version 08.0.10gT311
Boot-Monitor Image size = 786944, Version:10.1.05T310 (kxz10105)
HW: Stackable ICX6450-24-HPOE
==========================================================================
UNIT 1: SL 1: ICX6450-24p POE 24-port Management Module
Serial #: XXXXXXXXXXX
License: BASE_SOFT_PACKAGE (LID: dbtIHHHnFFl)
P-ENGINE 0: type DEF0, rev 01
==========================================================================
UNIT 1: SL 2: ICX6450-SFP-Plus 4port 40G Module
==========================================================================
800 MHz ARM processor ARMv5TE, 400 MHz bus
65536 KB flash memory
512 MB DRAM
STACKID 1 system uptime is 22 minutes 26 seconds
The system : started=warm start reloaded=by "reload"

ICX6450-24P Switch>65536 KB flash memory
Invalid input -> 65536 KB flash memory
Type ? for a list
ICX6450-24P Switch>
 

CED6688

New Member
Dec 4, 2019
15
10
3
For those of you who commented that you wanted the BGP to work with Kubernetes/Calico, I'm finding that the trivial solution is to just run Quagga on one or more of your nodes. I originally tried this with BIRD, but had a few issues due to bird already running in the calico-node container (possible to work around, but inconvenient).

I just pulled the version of Quagga from the Ubuntu (18.04) repository and have it reflecting the BGP routes put into the kernel routing table by Calico. You'll need to filter the table as I show below or bad things will happen, but once in place, this is zero maintenance and completely dynamic.

Here is an example Quagga configuration file that I put in /etc/quagga/osfpd.conf:
Code:
... stock configuration / interfaces / etc ...
router ospf
 ospf router-id 10.255.255.13 <-- Set this to a loopback address on the local box
 redistribute kernel !<-- This is the magic...but you MUST use a distribute-list + access-list or bad things will happen
 network 192.168.2.0/24 area 1.1.1.1 !<-- if you don't have another area, set this to area 0... set this to the network where you share an interface with your router
 distribute-list calico out kernel !<-- tells it to use the calico filter below to only redistribute routes from the calico subnets
!
access-list calico permit 10.244.0.0/16 !<-- this is my Cluster's CIDR
access-list calico permit 10.96.0.0/12 !<-- this is the Services CIDR, though I currently don't distribute this to BGP
access-list calico permit 10.255.255.13/32 !<-- this is a local loopback address
access-list calico deny any
On Ubuntu 18.04 w/ netplan, you can just add a line to your netplan file in /etc/netplan
Code:
network:
    ethernets:
        lo:
            addresses: [10.255.255.13/32]
On the ICX6450 (or any of the other routers we're talking about here), this couldn't be easier...
Code:
router ospf
 area 0 
 area z (whatever you set your area to above...)

interface ve xxx (the ve on the VLAN from the ospfd.conf file above)
 ip ospf area z
If you have full-mesh BGP enabled in calico's BGP config (the default), doing this on a single node will result in your ICX being dynamically updated with all of the routes needed to fully route your Kubernetes network. You can turn NAT off and create ACLs, etc... on your network devices. If you turn off full-mesh, you need to run quagga ospfd on each of your Kubernetes cluster nodes. I personally run this on two nodes for HA.

It isn't using the quagga node as transit...it's simply announcing the routes, so traffic goes directly to the correct node as you would expect/want.
 
  • Like
Reactions: groove

dashpuppy

Member
Dec 16, 2018
48
21
8
Can anyone provide links to SFP+ transceivers at a low price? The best I've found so far is: Cooper 30 m SFP+ Transceiver: Brocade 10G-SFPP-T

Also, what do I miss without the optical monitoring and would all these 'brocade compatible' transceivers miss that?

I just bought some Fiber ones and a 150' fiber cable for my test setup, SFP module's that are 10G were 30$ each, but you want copper right ?
 

fohdeesha

Kaini Industries
Nov 20, 2016
2,727
3,075
113
33
fohdeesha.com
so uhhh, I just booted Netberg Aurora 220 firmware on the ICX7250

Code:
 UBOOT: BE Mode
VENENO2 48 Copper (Non-POE), PVT1
SYS CPLD VER: 0x6, Released Ver: 0
I2C Read Failure, i 0 [ret 0x1], addr 0x52, muxdata 0x0, bus 0, len 11
EEPROM is not TLV format or CRC32 error.
I2C Read Failure, i 0 [ret 0x1], addr 0x52, muxdata 0x0, bus 0, len 11
I2C Read Failure, i 0 [ret 0x1], addr 0x52, muxdata 0x0, bus 0, len 10
 Signature mismatch Writing Shmoo values into flash .....
SF: Got idcodes
00000000: c2 20 17 c2 20    . ..
et0: gmac_serdes_init() PLL ready brought up exit
 
SF: Got idcodes
00000000: c2 20 17 c2 20    . ..
========== relocate address: 0xdfd1c000, offset 0xc1d1c000 ==========
Hit any key to stop autoboot:  0
DIABLO>
Netberg Aurora 220

why I did this, I have no idea
 

vangoose

Active Member
May 21, 2019
326
104
43
Canada
  • Like
Reactions: dashpuppy

eduncan911

The New James Dean
Jul 27, 2015
648
506
93
eduncan911.com
Yeah, many consumer motherboards are starting to come with 2.5 and 10 Gbps RJ45 10Gbase-T "copper" onboard. My last two desktop mobos had them.

And now my AMD server (Threadripper Gen2) has one.

Not to mention my USB-C thunderbolt 3 docking station with dual 4k, has a 10 Gbps 10GbaseT!

Lastly, some mainstream routers with WiFi6 now come with 2.5 or 5 Gbps RJ45s to support that new standard and bandwidth (why they don't go straight to 10Gbps, I don't know...).

I know, these Brocade switches are only 1/10 Gbps. But still, excluding the $800-Asus-routers-I'll-never-buy-anyways, almost everything I use daily has 10 Gbps copper RJ45.

Hence, the SPF+ copper adapters can really be useful these days.
 
  • Like
Reactions: dashpuppy