Brocade ICX Series (cheap & powerful 10gbE/40gbE switching)
- Thread starter fohdeesha
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
Well, I'd deleted my post because you'd followed up while I was writing it, and got it to work which, um, OK, makes it all the stranger. They could have just moved the USB connector to the back, which is completely blank on these, used a stacked pair of rj45s in the front and any typical network guy could just walk in and do what he needed to do with what he had with him. I assume that USB is for save/restores and firmware flashes. Which in most cases can usually be done another way. The 6610s don't have a USB of any sort so I assume this is somehow something to do with the 7250s being Linux based and the 6610s not. But anyway, seems like a lot of shenanigans to solve a problem, in an obscure way, that I'm not convinced was really a problem.
Anyway, watched that Youtube video, and on the USB-C type switches, they seem to work as you'd expect it to work. The USB port is actually USB and the RJ45 is for a typical Cisco type console cable, either way, it is a perfectly reasonable solution. In the mini USB switches, apparently, not so much, no RJ45 at all, and the mini USB isn't really a mini USB, it's an RS232 port wired onto mini USB connector. Making that mini USB actually a USB port would have been a better cheaper solution, no custom cable required and more options for OTS solutions, or move the USB-A port to the back on the 48 port version of the switch and over on the 24 port version, and put stacked RJ45s there and it's a typical Cisco cable config, also cheaper to produce. The included cable doesn't make it make any sense, it just makes what they did happen to work.
Anyway, glad to see you got it to work, and will likely be referring to your posts soon as I have a pair of ebay 7250s enroute, and pretty much the same stuff lying around. Knowing what I have learned from your posts, I think that hypothetical argument with the Brocade team would start with; What on earth are you thinking? Because their side of the argument only makes any sense after the fact. You'd have to have already committed to having those cables made to be included. If you are arguing with them before they have done so, it's nonsense.
Over the last several days I have looked at many tech docs about the 7250, and many many ebay listings for those switches, I don't recall any tech docs mentioning that obscure console cable or saw it included in any ebay listing. I've gone through fohdeesha's guides to make sure it'll do what I need it to do, but didn't see where he even explained that. Anyway good to know.
It is explained in the "Technical Specification" and in the "Installation Guide" of the ICX 7250. It is not one single cable but a set of two.
Years ago, when I I bought my 7250, it was hard to get and verry expansive. So I decided to build my own.
As far as console cables; also consider that people sometimes confuse the console port with an ethernet port, e.g. on a 6450 or 6610. Could have been a reason to have a console port that can't be confused with an other port. mini usb was pretty big at the time but was quickly replaced with micro usb.
Brocade L3 functionality is determined by the type of firmware version you are using, not by a license you install. There is a "switch" firmware and a "router" firmware. In order to get L3 functionality, you need to install the router firmware. You can still use the switch as a "layer 2" device if you install the router firmware, but it has the L3 functionality that isn't available on the plain "switch" firmware.
If the CLI prompt shows something like "ICX7250-24 Router" , then you are already using the "router" firmware. If it only shows as "ICX7250-24" or ICX7250-24 Switch", then you know you are using the switch firmware and will need to upload the router firmware to get access to the L3 functionality.
Last edited:
Here is a way to clearly identify the firmware:
The R in the firmware stand for Router and the S in the Firmware stands for Switch.
You can download the router version on the Ruckus homepage and flash it.
This Example is the router firmware: SPR09010k.bin
Code:
SSH@switch01>show version
Copyright (c) Ruckus Networks, Inc. All rights reserved.
UNIT 1: compiled on Jun 27 2025 at 08:27:18 labeled as SPR09010k
(33554432 bytes) from Primary SPR09010k.bin (UFI)
SW: Version 09.0.10kT213
Compressed Primary Boot Code size = 786944, Version:10.1.27T225 (mnz10127)
Compiled on Fri Feb 7 06:30:52 2025
HW: Stackable ICX7150-C12-POE
==========================================================================
UNIT 1: SL 1: ICX7150-C12-2X1G POE 12-port Management Module
Serial #:FEK3848P063
Software Package: ICX7150_L3_SOFT_PACKAGE
Current License: 2X10GR
P-ASIC 0: type B160, rev 11 Chip BCM56160_B0
==========================================================================
UNIT 1: SL 2: ICX7150-2X1GC 2-port 2G Module
==========================================================================
UNIT 1: SL 3: ICX7150-2X10GF 2-port 20G Module
==========================================================================
1000 MHz ARM processor ARMv7 88 MHz bus
8 MB boot flash memory
2 GB code flash memory
1 GB DRAM
STACKID 1 system uptime is 59 day(s) 9 hour(s) 20 minute(s) 4 second(s)
The system started at 09:07:08 GMT+01 Fri Nov 14 2025
The system : started=cold startThank you both. The FCX switch I've got is using a router firmware as far as I can tell. It has "Router" in its default hostname and is using the firmware from these instructions which does have the R in its filename ICX6610-FCX/FCXR08030u.bin. But it's still complaining about the missing L3 license and L3 related commands are missing
what commands are complaining about a missing license? There's an "ADV" license which unlocks some advanced l3 stuff, but it's only VRFs and tunnels basically. everything else is included in the base system. IUf you need the VRFs and tunnels, follow the i2c licensing instruction at the end of the FCX guide as @klui said
Hello everyone,
I just got a 6430-24p and followed Fohdeeshas guide for new firmware and initial setup. The reason I chose the 6430-24p over the others is because of it's lower power consumption while still providing POE+, however I'm seeing more than double the idle power consumption listed in the data sheet.
The data sheet says 9w idle, but I'm seeing 20.8w idle with no ports connected, and 21.4w idle with one port connected. I read through a lot of wattage related posts in this thread and many people claim they get similar or identical figures compared to the data sheet for other 64xx models, so I came to ask if I am missing something.
My meter is a kill-a-watt style wall meter, but a different brand. I tested it several times against other meters, and with a known load, and it's accurate enough, within 1 watt or less. I let the switch run for over half an hour after following Fohdeeshas instructions including the factory reset, unplugged it and plugged it back in, let it sit some more, and it still didn't change.
Any ideas on how to get lower power consumption, or why I'm seeing double the spec'd idle figures?
I just got a 6430-24p and followed Fohdeeshas guide for new firmware and initial setup. The reason I chose the 6430-24p over the others is because of it's lower power consumption while still providing POE+, however I'm seeing more than double the idle power consumption listed in the data sheet.
The data sheet says 9w idle, but I'm seeing 20.8w idle with no ports connected, and 21.4w idle with one port connected. I read through a lot of wattage related posts in this thread and many people claim they get similar or identical figures compared to the data sheet for other 64xx models, so I came to ask if I am missing something.
My meter is a kill-a-watt style wall meter, but a different brand. I tested it several times against other meters, and with a known load, and it's accurate enough, within 1 watt or less. I let the switch run for over half an hour after following Fohdeeshas instructions including the factory reset, unplugged it and plugged it back in, let it sit some more, and it still didn't change.
Any ideas on how to get lower power consumption, or why I'm seeing double the spec'd idle figures?
I just thought I'd post this in case anyone is interested about the airflow direction in the ICX-6610/6650 power supplies and fan modules:
Both the fan modules and the power supplies of the 250W/non-PoE switches use the same fan model numbers for each respective module, just installed with the airflow in opposite directions. The fan modules use a pair of Delta FFB0412UHN (.81A) and the Power Supplies use a Sanyo-Denki SanAce40 9GV0412P3J03 (6.0A).
So I don't see any reason why you can't just flip the fans to reverse the airflow directions in these switches. I don't have both versions of the 1000W/PoE Supplies to verify that this is also true there but I assume it is. But I think converting a back to front airflow switch to front to back is the more likely desirable scenario for home lab. Generally speaking most surplus non-PoE are back to front airflow, as they were typically installed as top of rack in the rear of server racks, whereas PoE switches were more typically installed as office distro switches and are almost all front to back airflow. Meaning almost all 1000W supplies are "-I" type supplies and fans and almost all 250W supplies are "-E" supplies and fans.
Anyway, I am sure I am not the only one who may want to have a full power PoE switch in the same rack as a more economical to run non-PoE version.
Both the fan modules and the power supplies of the 250W/non-PoE switches use the same fan model numbers for each respective module, just installed with the airflow in opposite directions. The fan modules use a pair of Delta FFB0412UHN (.81A) and the Power Supplies use a Sanyo-Denki SanAce40 9GV0412P3J03 (6.0A).
So I don't see any reason why you can't just flip the fans to reverse the airflow directions in these switches. I don't have both versions of the 1000W/PoE Supplies to verify that this is also true there but I assume it is. But I think converting a back to front airflow switch to front to back is the more likely desirable scenario for home lab. Generally speaking most surplus non-PoE are back to front airflow, as they were typically installed as top of rack in the rear of server racks, whereas PoE switches were more typically installed as office distro switches and are almost all front to back airflow. Meaning almost all 1000W supplies are "-I" type supplies and fans and almost all 250W supplies are "-E" supplies and fans.
Anyway, I am sure I am not the only one who may want to have a full power PoE switch in the same rack as a more economical to run non-PoE version.
As this thread has 548 pages in it and the search function didn't help me narrow down what I'm looking for, I thought I might ask.
I have two ICX-6450s set up in a stack with two EPS1500s attached to them. I'm getting SNMP info from them but I'm not 100% sure of which OID has information from the second unit for things like fan and power supply status. I'm getting that info but I think just from Unit 1. Also, I was wondering if there's any status information available from the ESP1500s as well such as temp.
So, any information on OIDs and data gathering into Prometheus would be very welcome.
Additionally, I believe the fans in the EPS1500s are exactly the same make and model as the ones I replaced in the 6450, but I just wanted to confirm before I replace them as well.
For context I am running a bunch of PoE ESP32 sensors around my house so the PoE load isn't massive, but given the cost of the units is so low I figured why not add the EPS1500s?
I have two ICX-6450s set up in a stack with two EPS1500s attached to them. I'm getting SNMP info from them but I'm not 100% sure of which OID has information from the second unit for things like fan and power supply status. I'm getting that info but I think just from Unit 1. Also, I was wondering if there's any status information available from the ESP1500s as well such as temp.
So, any information on OIDs and data gathering into Prometheus would be very welcome.
Additionally, I believe the fans in the EPS1500s are exactly the same make and model as the ones I replaced in the 6450, but I just wanted to confirm before I replace them as well.
For context I am running a bunch of PoE ESP32 sensors around my house so the PoE load isn't massive, but given the cost of the units is so low I figured why not add the EPS1500s?
I got bitten by the Brocade bug and got a ICX7250-48 on ebay. I have been able to set up tftp and follow the docs at ICX7250 / ICX7450 - Fohdeesha Docs up to "First Boot and Login".
Unfortunately, this is how far I get, then the switch reboots after a while and repeats the whole sequence again.
Is this recoverable or should I just try and get a refund because the switch is clearly unsalvageable?
Code:
Ruckus Networks Bootloader: 10.1.18T215 (Jul 13 2020 - 01:53:05)
Validate Shmoo parameters stored in flash ..... OK
Restoring Shmoo parameters from flash .....
Running simple memory test ..... OK
ICX7250-48 Copper (POE), PVT1
SYS CPLD VER: 0x6, Released Ver: 0
Enter 'b' to stop at boot monitor: 3 ... 2 ... 1 ... 0
bootdelay: ===
Booting image from Primary
NAND read: device 0 offset 0x0, size 0x2000000
................................ 33554432 bytes read: OK
## Booting kernel from FIT Image at 70000200 ...
Using 'conf@3' configuration
Trying 'kernel@1' kernel subimage
Description: Ruckus Linux SP/SI
Type: Kernel Image
Compression: lzma compressed
Data Start: 0x700002d0
Data Size: 2422888 Bytes = Architecture: ARM
OS: Linux
Load Address: 0x61008000
Entry Point: 0x61008000
Hash algo: crc32
Hash value: 2a818577
Verifying Hash Integrity ... crc32+ OK
## Loading init Ramdisk from FIT Image at 70000200 ...
Using 'conf@3' configuration
Trying 'ramdisk@1' ramdisk subimage
Description: Ramdisk -rootfs
Type: RAMDisk Image
Compression: lzma compressed
Data Start: 0x704a06b0
Data Size: 27720752 Bytes = Architecture: ARM
OS: Linux
Load Address: 0x00000000
Entry Point: 0x00000000
Hash algo: crc32
Hash value: e9b25165
Verifying Hash Integrity ... crc32+ OK
## Flattened Device Tree from FIT Image at 70000200
Using 'conf@3' configuration
Trying 'fdt@2' FDT blob subimage
Description: Flattened Device Tree blob
Type: Flat Device Tree
Compression: uncompressed
Data Start: 0x7049c17c
Data Size: 8730 Bytes = Architecture: ARM
Hash algo: crc32
Hash value: a1390d0f
Verifying Hash Integrity ... crc32+ OK
Loading FDT from 0x7049c17c to 0x65008000
Booting using the fdt blob at 0x65008000
Uncompressing Kernel Image ... OK
Using Device Tree in place at 65008000, end 6500d219
Starting kernel ...
bcmiproc-gmac 18023000.ethernet: can't find MAC address
Mounting file system...
booting with 4.4 kernel
PLATFORM SI
dd cmd: UBI
Config partition mounted
Resources folder check done!
Please download UFI image to install packages
/etc/process_packages.sh: line 148: can't create /.pkg/error.log: nonexistent directory
Package installation is incomplete, related functionalities may not be available
dhcpd executable file is not present.
Starting up system statistics daemon...
Starting systatsd done
Starting up health monitor daemon...
Log Level 7: Daemonized systatsd
Systatsd Logging Env is not set, initializing loglevel to: 7
Setting connection to systemlogger
Starting hmond done
Hmond Logging Env is not set, initializing loglelvel to: I
Setting connection to systemlogger
Skip the license and configuraiton backup (no changes).
Kernel crashdump is disabled.
Skip the TPM certificates backup (no changes).
FIPS reset is enable.
Removing startup config & backup
Removing stacking.boot
Remodma_mem_base: 0x96000000, dma_mem_len: 0x8000000, warm_mem_base: 0xdf500000, warm_mem_len: 0xb00000
ving SSL_CERT_FILE
Removing SSH_RSA_HOSTKEY_FILE
Removing SSH_DSA_HOSTKEY_FILE
Creating TUN device
Enabling time stamp.
*** Crash dump information is available, please collect it using "supportsave" CLI ***
Starting the FastIron.
USER=root
boot_partition=pri_partition
HOME=/
GCOV_PREFIX=/fast_iron
ethaddr=78:a6:e1:1a:38:fb?
TERM=vt102
PATH=/sbin:/usr/sbin:/bin:/usr/bin
ip=:192.168.178.67:10.14.234.1:255.255.255.0:SI:eth0:off
SHELL=/bin/sh
PWD=/
[sp-gpio]:sil_gpiochip_register: gpiochip registeration
i2c-0.smbus .iproc-smb-1 .SMBus adapter.3
i2c-1.smbus .iproc-smb-1 .SMBus adapter.3
FIPS Disabled:PORT NOT DISABLED
platform type 84
OS>
Loaded Image SPR08090mc from Pri Built on (UTC): Mon Oct 4 17:51:11 2021
Enabling Console LoggingApplying factory defaults..
INFO: startup config data is not available, try to read from backup
INFO: startup config data in the backup area is not available
Hotplugger Daemon Initialized ..
set_board_level: gi_board_type = 111HWM: Error writing value:0x69 in offset 0x3d
Failed to get mux lock for bus 0
Stack trace of software that has the lock:
stack: 000551a0 00502f84 0047d8c0 0047dab0 0048cd64 00463240 0046a8e0 004630a0 004e88ec 003526bc 00342484 00333550 004b87a4 b6e76a20 b6f76195Is this recoverable or should I just try and get a refund because the switch is clearly unsalvageable?
I am putting together an ICX6610 stack of several switches, I want to use the 40G QSFP+ ports for this. Half of those ports are apparently permanently configured as native 40G ports and half of them are 4x10G ports. My understanding is that passive DACs don't really care about one type vs the other, so I can just use inexpensive passive DAC cables if they are within 5m of each other, right? But I'll have one segment of the stack that isn't within 5m, it's more like 10m. So I have 2 choices there, use AOC cables for that segment or tranceivers and fiber, which is probably the way I'd prefer to do it, because I kinda get the impression that a lot of these AOC cables are really just tranceivers and fiber cables permently assembled together, more or less.
When I look at 40G tranceivers on fs.com though, they seem to differentiate between 40G tranceivers (QSFP-40G-SR4-S) and 4x10G breakout capable tranceivers (QSFP-40G-SR4) with a not so insignificant cost difference (about double). But my question is do you need to use a breakout capable tranceiver if I am not actually breaking anything out? I want to use these back to back. Or are there any recommondations for cheap tranceivers that just work in either port type? Used ebay surplus is fine. And what cable do I need between them, if they are just back to back? A, B, or C, key up, key down? I've never used MTP/MPO cables before, so kinda new to this.
When I look at 40G tranceivers on fs.com though, they seem to differentiate between 40G tranceivers (QSFP-40G-SR4-S) and 4x10G breakout capable tranceivers (QSFP-40G-SR4) with a not so insignificant cost difference (about double). But my question is do you need to use a breakout capable tranceiver if I am not actually breaking anything out? I want to use these back to back. Or are there any recommondations for cheap tranceivers that just work in either port type? Used ebay surplus is fine. And what cable do I need between them, if they are just back to back? A, B, or C, key up, key down? I've never used MTP/MPO cables before, so kinda new to this.