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.

RoachedCoach

Member
Feb 4, 2020
46
54
18
I've got an ICX 7250-48 that I'm doing the fan mod on based on a comment way on this forum. I had also added the fan to the CPU like they recommended, but the original poster stated that he had spliced into one of the fans for power. Now the 7250-48 comes with 2 fans, but has the headers for the 3rd. It looks like it should work if I can enable the 3rd fan, but I don't know how in my current firmware (08.0.95). Is it even possible in this firmware version or would I need to upgrade to 09.0.10? I'm really hoping that it isn't something entirely disabled in the firmware.
I think this was me. If i were you, I'd just plug one of the fans you're replacing into the 3rd header and fire it up and see if it runs all 3. I suspect that the third one is probably enabled on all models, regardless of firmware settings.

But if it doesn't work, you could splice into the second fan for the CPU fan and not even worry about fan 3.
 
  • Like
Reactions: bgubs

thehedgefrog

New Member
Sep 14, 2022
8
1
3
I've got an ICX 7250-48 that I'm doing the fan mod on based on a comment way on this forum. I had also added the fan to the CPU like they recommended, but the original poster stated that he had spliced into one of the fans for power. Now the 7250-48 comes with 2 fans, but has the headers for the 3rd. It looks like it should work if I can enable the 3rd fan, but I don't know how in my current firmware (08.0.95). Is it even possible in this firmware version or would I need to upgrade to 09.0.10? I'm really hoping that it isn't something entirely disabled in the firmware.
You can test with a multimeter. If that doesn't work, you can use a splitter and power 2 fans off of one header.
 

i386

Well-Known Member
Mar 18, 2016
4,527
1,696
113
35
Germany
Is a 7150-48ZP in the E8X10GR sku (with all licenses/features and protocols enabled) for 666£ (~1080€) a good deal? :D
 

bgubs

New Member
Jul 29, 2024
4
2
3
I think this was me. If i were you, I'd just plug one of the fans you're replacing into the 3rd header and fire it up and see if it runs all 3. I suspect that the third one is probably enabled on all models, regardless of firmware settings.

But if it doesn't work, you could splice into the second fan for the CPU fan and not even worry about fan 3.
Oh, hey! It is you! I bought the exact fans you recommended and they have definitely had a noticeable reduction in temps/noise.

I took the switch apart again. Using my multimeter, I confirmed power does not get to the 3rd port. There's two maybe 100uF capacitors and a Mosfet or some NPN transistor that's missing to handle switching the fan on/off. Could definitely run a wire with power to the port, but I think I'll do the more sensible thing and just use a new connector to split one of the existing connections like recommended by you before.
 
  • Like
Reactions: RoachedCoach

bgubs

New Member
Jul 29, 2024
4
2
3
Wired up the other fan inline with fan #2 using a new connector. It works great. I'm now idling around 46-47C. I wish I had a benchmark to go off of for when I hook my network back up, then I could've really compared the difference in temps.

I'm still curious though about fan port #3. I'm wondering if I found the 3 components that were missing and soldered them on to the board whether or not the fan would actually work (with PWM too) and recognized in `show chassis`. Absolutely a waste of money and time regardless, but sounds like a fun and neat experience.
 
  • Like
Reactions: RoachedCoach

bugeater4

New Member
Aug 2, 2024
2
0
1
Im reading through this 500 page forum but cant seem to see anyone having this issue. Im stacking 2 ruckus switches. both are ICX7150 but one is a 24P and one is a 48P. we have verified licencing on the sfp ports and have FS brand 10g DAC cables. both are on firmware version 8.0.90jT211 and can boot up but when the unit 2 decides become the standby on the stack, it triggers a reboot causing the second switch (unit 2, the 48P) to reboot and constantly has a blinking amber light for the SYS light. anyone know what I can do to fix this?
 

bugeater4

New Member
Aug 2, 2024
2
0
1
have you considered posting the configs from each switch ? :eek:
there are no changes to the configs outside of the initial default config as we have wiped them over and over, for security reasons as this is for a company, I can not provide an entire config for both
 

NablaSquaredG

Bringing 100G switches to homelabs
Aug 17, 2020
1,684
1,116
113
Im reading through this 500 page forum but cant seem to see anyone having this issue. Im stacking 2 ruckus switches. both are ICX7150 but one is a 24P and one is a 48P. we have verified licencing on the sfp ports and have FS brand 10g DAC cables. both are on firmware version 8.0.90jT211 and can boot up but when the unit 2 decides become the standby on the stack, it triggers a reboot causing the second switch (unit 2, the 48P) to reboot and constantly has a blinking amber light for the SYS light. anyone know what I can do to fix this?
Have you considered upgrading to 08.0.95p? And if that does not fix the issue, write a post in the Ruckus forums?
 
  • Like
Reactions: jode

kpfleming

Active Member
Dec 28, 2021
432
222
43
Pelham NY USA
there are no changes to the configs outside of the initial default config as we have wiped them over and over, for security reasons as this is for a company, I can not provide an entire config for both
Have you explicitly cleared the 'stacking' configuration? It's stored separately, to some degree, from the regular configuration file. That would be done using the 'stack unconfigure' command.
 
  • Like
Reactions: Jason Antes

clcorbin

Member
Feb 15, 2014
76
10
8
I have a pair of ICX 6610s in a stack that have been very happily running for three years or so. Major overkill for a home network, but lots of fun!

Currently, one switch is located in my server rack and the other is out in the garage where most of the homes original networking terminates. The switches are stacked with the stacking ports 2/2/1 and 2/2/6 via MPO fiber. Both running latest firmware, have the handful of vlans configured, etc.

Now, I need to break the stack. I want to replace the server rack switch with a 7750-48F, mainly because I need more 10Gbe ports and more 40Gbe to support a 4 node proxmox cluster I am setting up (each nodes has dual 10Gbe ports for admin and VMs and dual 40Gbe ports for cluster). The 7750 will replace the 6610 and a Mellanox SX6036.

My question is this: What is the best approach for this? My thinking is the first step would be to disable stacking and revert the switches back to individual switches with dual 40Gbe LAGs through the current stacking ports. Once that is reconfigured and working, then I can configure the 7750 ports as needed to match my current network and swap it in for the 6610.

What is the best way of going about this to minimize the possibility of hosing everything I have spent three years fine tuning (with MUCH help from this forum and this thread!)?

Thanks!

Clint
 

tls

New Member
Oct 26, 2022
4
0
1
I have a mixed network of ICX and Cisco switches, running MSTP. I'd like to separate it into multiple MSTP regions, which in a pure Cisco configuration I would do with a simple:
Code:
spanning-tree mst configuration
name whatever
Using a different name for each region. But I can't find any command at all to set the region name in FastIron. I see the mstp name command, but from the documentation synopsis it makes no sense, claiming that it is a name "for the device", and that "the name applies to the switch that can have many different VLANs that can belong to many different MSTP regions". This makes no sense - names in MSTP are per-region, not per-switch. Is the documentation just wrong? It gives this example, too, which sure looks like whoever wrote it thought this was a "switch name" not a "region name" - but there is no other way I can find to set the region name!
Code:
device(config)# mstp name Device1
 

tls

New Member
Oct 26, 2022
4
0
1
I see the mstp name command, but from the documentation synopsis it makes no sense, claiming that it is a name "for the device", and that "the name applies to the switch that can have many different VLANs that can belong to many different MSTP regions". This makes no sense - names in MSTP are per-region, not per-switch. Is the documentation just wrong? It gives this example, too, which sure looks like whoever wrote it thought this was a "switch name" not a "region name" - but there is no other way I can find to set the region name!
Code:
device(config)# mstp name Device1
I did some guesswork and testing. The documentation (of course?) is wrong. mstp name sets the region name, just like on everyone else's switches; it does not set the "device name" which is nonsense from a MSTP point of view. The example is bogus and misleading, too, of course; it should probably be changed to use a region name of "mymst" since that's what's used in all of Cisco's configuration examples, or something more clear like "Region1".

Watch out for the MSTP "revision" also - FastIron defaults this to 0 like most other switches, but there are a few brands and product lines out there that default it to 1. If both name and revision don't match, you will end up with multiple MST regions you did not design, all communicating by RSTP or even old-school STP across the "boundaries", and are likely to lose redundancy you designed into your network, since links beyond boundaries are not visible to each accidentally configured region's root bridge.
 

ProxmoxProphet

New Member
Apr 2, 2024
8
0
1
Earth
I bought an ICX7250-48P a few months ago, and, other than some strange IPv6 issues that resolved once I upgraded to 09.0.10 it's been great, but recently I noticed another frustrating issue that I don't know how to fix and I hope someone who encountered this issue can reply with how they fixed it.

I noticed a problem recently with the switch not being able to handle wifi devices roaming between different APs after moving to a new house where I had to buy multiple APs to get connectivity in the whole house (my old house was smaller so I had only 1 AP and therefore never encountered this problem). So here is an explanation of how the problem presents itself: Let's say I am downstairs, connected to my home wifi via the downstairs AP on my phone, I can access the internet just fine, I can access locally-hosted services like nextcloud, jellyfin, etc just fine BUT when I go upstairs to my bedroom, and my phone roams from the downstairs AP to the AP in the upstairs hallway, now the phone can no longer reach anything that is more than one hop away. It can reach other devices in the same VLAN just fine, it can reach other wifi devices, the switch on that VLAN, etc, but it can no longer reach the internet or any locally-hosted services (because the locally-hosted servers are on a different VLAN, and the uplink from the switch to my firewall is also on its own VLAN, so its 2 hops away to those). And if I statically assign the phone a different IP address, then all of a sudden now it can reach the internet and locally-hosted services again (until it roams to a different AP once again and the problem happens again). And this issue is not just with my phone but with any Wi-Fi connected device

Also, another problem I noticed a while ago as well that frustrates me: Sometimes, when a device on the network is assigned an IP that was previously leased to some other device by the DHCP server, then that device cannot reach the internet or anything else thats more than one hop away unless I give it a different IP. I'm not sure but maybe these two issues are related, because the problematic behavior (not being able to reach anything more than 1 hop away) is present in both


Here is a brief description of my network topology in case it would help anyone understand something better:

ISP fiber cable comes into my home and connects to the ONT, then the ONT connects to my opnsense firewall with ethernet. From there, my opnsense firewall connects to a 2.5G/10G managed switch via SFP+ DAC. The 2.5G ports on the switch are used to connect downstairs AP and other things on the user VLAN, and the other SFP+ port on that switch is connected to a singlemode fiber cable which runs to the upstairs hallway closet where my ICX7250 is. This cable carries userVLAN traffic from downstairs to the ICX7250 and the traffic for the transit VLAN between the ICX7250 and Opnsense is also sent down this cable. Then, there are two more APs upstairs connected to the ICX7250 (on the same VLAN as the downstairs APs) as well as some other servers and PCs, a fiber cable that connects to a Mikrotik 10G switch on the third floor, and other things connected to the ICX7250 as well.
 
Last edited:

jode

Member
Jul 27, 2021
65
57
18
I noticed a problem recently with the switch not being able to handle wifi devices roaming between different APs after moving to a new house where I had to buy multiple APs to get connectivity in the whole house (my old house was smaller so I had only 1 AP and therefore never encountered this problem).
I have quite a similar setup which works fine on a ICX-6450. So, I don't think it's necessarily a switch limitation.

In my setup, I can see devices moving from one port (AP1) to another port (AP2) in the log.
Try sh log and look for that activity.

I would expect some configuration issue, possibly on the 2.5G/10G switch (wild guess, based on my setup experience). Make sure that all VLANs you use with your APs are connected via both switches.
 
Last edited:

bgubs

New Member
Jul 29, 2024
4
2
3
Wired up the other fan inline with fan #2 using a new connector. It works great. I'm now idling around 46-47C. I wish I had a benchmark to go off of for when I hook my network back up, then I could've really compared the difference in temps.

I'm still curious though about fan port #3. I'm wondering if I found the 3 components that were missing and soldered them on to the board whether or not the fan would actually work (with PWM too) and recognized in `show chassis`. Absolutely a waste of money and time regardless, but sounds like a fun and neat experience.
Well experiment was a partial bust. I was able to find a similar mosfet and had some capacitors on hand. Soldered them in today and the fan will turn on, but seems there's no logic sent to the 3rd pin. I'm going to trace a little further up the line, but I think unless someone with more know-how is able to enable the 3rd pin in the firmware the other half of the experiment is over. Oh well, was a fun test and worth a try.
 

ProxmoxProphet

New Member
Apr 2, 2024
8
0
1
Earth
I have quite a similar setup which works fine on a ICX-6450. So, I don't think it's necessarily a switch limitation.

In my setup, I can see devices moving from one port (AP1) to another port (AP2) in the log.
Try sh log and look for that activity.

I would expect some configuration issue, possibly on the 2.5G/10G switch (wild guess, based on my setup experience). Make sure that all VLANs you use with your APs are connected via both switches.
Hi, thanks for sharing your experience, it's good to hear that you don't have this problem, but also strange, I'm not sure what is causing this problem for me.

I walked with my phone to downstairs and back (confirming in the TP link app that the phone did actually roam to the downstairs AP and back to the office AP) and then I ran that command, but I didn't see any output about my phone, or any other devices, roaming between APs. The only output was stuff about failed and successful SSH logins, some stuff about STP, and some stuff about the switch disabling PoE on certain ports because a non-PoE devices was connected, but nothing about AP roaming.

I also thought that the 2.5G/10G switch may be responsible for the issues, but I checked the configuration on that and the appropriate VLANs were passed, the ethernet ports on that switch are all untagged for my User VLAN (VLAN 110) and the SFP+ ports are both tagged for the User VLAN and the Transit VLAN.

Also, I also notice this issue not just when roaming between downstairs and second floor APs, but also when roaming between the second floor APs and the AP on the third floor in my office, and the 2.5G/10G switch is on the first floor, so I don't think that 2.5/10G switch would be involved when roaming between second and third floor APs. Also I checked the configuration on the Mikrotik switch on the 3rd floor (which connects to the 3rd floor AP) and the VLANs were configured appropriately on there as well
 

fohdeesha

Kaini Industries
Nov 20, 2016
2,868
3,368
113
34
fohdeesha.com
I bought an ICX7250-48P a few months ago, and, other than some strange IPv6 issues that resolved once I upgraded to 09.0.10 it's been great, but recently I noticed another frustrating issue that I don't know how to fix and I hope someone who encountered this issue can reply with how they fixed it.

I noticed a problem recently with the switch not being able to handle wifi devices roaming between different APs after moving to a new house where I had to buy multiple APs to get connectivity in the whole house (my old house was smaller so I had only 1 AP and therefore never encountered this problem). So here is an explanation of how the problem presents itself: Let's say I am downstairs, connected to my home wifi via the downstairs AP on my phone, I can access the internet just fine, I can access locally-hosted services like nextcloud, jellyfin, etc just fine BUT when I go upstairs to my bedroom, and my phone roams from the downstairs AP to the AP in the upstairs hallway, now the phone can no longer reach anything that is more than one hop away. It can reach other devices in the same VLAN just fine, it can reach other wifi devices, the switch on that VLAN, etc, but it can no longer reach the internet or any locally-hosted services (because the locally-hosted servers are on a different VLAN, and the uplink from the switch to my firewall is also on its own VLAN, so its 2 hops away to those). And if I statically assign the phone a different IP address, then all of a sudden now it can reach the internet and locally-hosted services again (until it roams to a different AP once again and the problem happens again). And this issue is not just with my phone but with any Wi-Fi connected device

Also, another problem I noticed a while ago as well that frustrates me: Sometimes, when a device on the network is assigned an IP that was previously leased to some other device by the DHCP server, then that device cannot reach the internet or anything else thats more than one hop away unless I give it a different IP. I'm not sure but maybe these two issues are related, because the problematic behavior (not being able to reach anything more than 1 hop away) is present in both


Here is a brief description of my network topology in case it would help anyone understand something better:

ISP fiber cable comes into my home and connects to the ONT, then the ONT connects to my opnsense firewall with ethernet. From there, my opnsense firewall connects to a 2.5G/10G managed switch via SFP+ DAC. The 2.5G ports on the switch are used to connect downstairs AP and other things on the user VLAN, and the other SFP+ port on that switch is connected to a singlemode fiber cable which runs to the upstairs hallway closet where my ICX7250 is. This cable carries userVLAN traffic from downstairs to the ICX7250 and the traffic for the transit VLAN between the ICX7250 and Opnsense is also sent down this cable. Then, there are two more APs upstairs connected to the ICX7250 (on the same VLAN as the downstairs APs) as well as some other servers and PCs, a fiber cable that connects to a Mikrotik 10G switch on the third floor, and other things connected to the ICX7250 as well.

this is an arp table issue most likely, have you changed anything weird in your config? previous fastiron versions had some bugs when it came to ipv6 ND, it wouldn't update the ND table when devices moved between physical switch ports, but that was fixed in like 8095 as far as I know
 
  • Like
Reactions: ProxmoxProphet

ProxmoxProphet

New Member
Apr 2, 2024
8
0
1
Earth
this is an arp table issue most likely, have you changed anything weird in your config? previous fastiron versions had some bugs when it came to ipv6 ND, it wouldn't update the ND table when devices moved between physical switch ports, but that was fixed in like 8095 as far as I know
Hi, I haven't changed my config recently, and I'm also on the version 09.0.10j_cd2T213 , so definitely newer than 8095
I made a comment in this thread a few months ago about an IPv6 issue that I resolved and my config in there is the same as my current config (with only minor changes like creating VLANs, adding VLANs to ports and such)

Is there any fix for the arp table issue you mentioned?