Turbocharge your Quanta LB6M - Flash To Brocade TurboIron - Big Update!

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

Corsaire

New Member
Feb 3, 2018
21
0
1
Hi there,

count one happy camper more :)
I did the procedure too and now have a "genuire" Brocade switch running home.

One thing I want to share here, I got an headache figuring how to make a 4-way LACP with a Netgear switch, so I want to share here my results.

From what I saw you need to set Timeout value on short on the concerned ports, both on your Netgear and the Brocade.
In Netgear GUI : Switching - LAG - Advanced - LACP Port Configuration

I used the four 1G ethernet ports

Code:
interface ethernet 25 ethernet 26 ethernet 27 ethernet 28
link-aggregate conf key 10000
link-aggregate conf timeout short
link-aggregate active

show link-aggregate

It takes a while to get up so give it time.
 
Last edited:

Corsaire

New Member
Feb 3, 2018
21
0
1
On a second note, temperature sensoring is a bitch...

On my modified switch it does takes temperature then stops doing so for no reason (I replaced all fans in there).
I had some really nice Cacti graphs...

Code:
>show chassis
Power supply 1 (NA - NA - Regular) present, status ok
Power supply 2 (NA - NA - Regular) present, status ok

Fan 1 ok, speed (auto): [[1]]<->2<->3
Fan 2 ok, speed (auto): [[1]]<->2<->3
Fan 3 ok, speed (auto): [[1]]<->2<->3

Fan controlled temperature: 0.0 deg-C

Fan speed switching temperature thresholds:
                Speed 1: NM<----->30                            deg-C
                Speed 2:       25<----->65                      deg-C
                Speed 3:             60<----->90                deg-C (shutdown)


Exhaust Side Temperature Readings:
        Current temperature : 0.0 deg-C
        Warning level.......: 80.0 deg-C
        Shutdown level......: 90.0 deg-C
Intake Side Temperature Readings:
        Current temperature : 0.0 deg-C
 

TheBloke

Active Member
Feb 23, 2017
200
40
28
44
Brighton, UK
On a second note, temperature sensoring is a bitch...

On my modified switch it does takes temperature then stops doing so for no reason (I replaced all fans in there).
I had some really nice Cacti graphs...

Code:
>show chassis
Power supply 1 (NA - NA - Regular) present, status ok
Power supply 2 (NA - NA - Regular) present, status ok

Fan 1 ok, speed (auto): [[1]]<->2<->3
Fan 2 ok, speed (auto): [[1]]<->2<->3
Fan 3 ok, speed (auto): [[1]]<->2<->3

Fan controlled temperature: 0.0 deg-C

Fan speed switching temperature thresholds:
                Speed 1: NM<----->30                            deg-C
                Speed 2:       25<----->65                      deg-C
                Speed 3:             60<----->90                deg-C (shutdown)


Exhaust Side Temperature Readings:
        Current temperature : 0.0 deg-C
        Warning level.......: 80.0 deg-C
        Shutdown level......: 90.0 deg-C
Intake Side Temperature Readings:
        Current temperature : 0.0 deg-C
I take it you're using 3-pin fans?

Read the last page of the original LB6M thread, starting with this post, for my many posts on this subject :) The temp sensing works with 4-pin PWM fans, or is retrievable using 3-pin fans with a messy hack involving a resistor on one of the fan connectors.
 

TheBloke

Active Member
Feb 23, 2017
200
40
28
44
Brighton, UK
Yes, for some of them.

The most infuriating part is that it worked fine until suddenly it decided to just say "0" everywhere...
Yup, exactly what happened to me. It worked for several reboots and a couple of hours of runtime following my installing 3 x 3-pin fans, then on the next power cycle I got no temps. During my testing I then did dozens of power cycles over the next 2-3 hours and it worked on two of them, both times being after I had first booted with a 4-pin fan in place then booted again with a 3-pin again. After that I never saw temps again with a 3-pin fan, until I found the hack solution.

Or rather, it's fan 3 (closest to PSU) that has to be 4-pin or have the resistor hack on it. If that fan is detected OK, it doesn't matter about the other two - they can be 3-pin, or not be there at all, and you still get temps.

I certainly don't understand why it worked for a bit then stopped. It's almost as if the fan/temp code has allowed the 'problem' to persist for a bit, then wrote something to flash indicating temp monitoring should be turned off in future if the proper fans aren't detected. Something crazy like that.

Anyway, it's fixable if you don't mind some messing about. I bought some 3 and 4-pin fan connectors with the intention of making a proper cable integrating the 120k resistor between the (empty) pin four and pin three. But then I ended up instead going for a single 12cm 4-pin PWM fan, so the hack proved unnecessary.

EDIT: Actually, you say 'some' of them? As above, if you have even a single 4-pin fan, if you connect it in fan position three (nearest to PSU; right-most, as you look from the back) you should be fine. At least in terms of getting temps, anyway - the PWM speed control isn't guaranteed to work right.
 
Last edited:

Corsaire

New Member
Feb 3, 2018
21
0
1
EDIT: Actually, you say 'some' of them? As above, if you have even a single 4-pin fan, if you connect it in fan position three (nearest to PSU; right-most, as you look from the back) you should be fine.
Now that we're talking about I have some doubts...
I made the fan modding a few months ago and i THINK I had at least one 4-pin fan in the lot, I would have to open the cover again to verify.. such a pain :p

If so I will just connect it to that connector.
 

TheBloke

Active Member
Feb 23, 2017
200
40
28
44
Brighton, UK
I made the fan modding a few months ago and i THINK I had at least one 4-pin fan in the lot, I would have to open the cover again to verify.. such a pain :p
Yeah, those 5 case screws can be pretty taxing ;)

So your temps have been working for months and just stopped working now? That would be quite different to my experience - unless of course you've not powered down the switch more than a couple of times in that time :)
 

Corsaire

New Member
Feb 3, 2018
21
0
1
Yeah, those 5 case screws can be pretty taxing ;)

So your temps have been working for months and just stopped working now? That would be quite different to my experience - unless of course you've not powered down the switch more than a couple of times in that time :)
Not months, more precisely :

I modded the fans, then got the temp with the SNMP OID.
At this moment I was still using the Quanta firmware.
Then suddenly it stopped showing and turned to "0"

This week-end I flashed the firmware and, surprise !
Using the cacti Brocade ICX template I saw that I had my temps back, even better with more probes but only two of them really active (if we could post pics directly on the forum I would have done so).
But yesterday without any reason (I even wasn't home) the temps started to show "0" ...
 

TheBloke

Active Member
Feb 23, 2017
200
40
28
44
Brighton, UK
OK interesting, so presumably there wasn't even a power cycle between when it worked and when it didn't.

The LB6M fan detection code is really weird, as documented at great length in my posts in the other thread. I did dozens of tests and I more or less know what works and what doesn't, but have little clue why it's like that.

The primary assumption is that the switch does some very specific tests on the fans, that replacement fans don't pass, and they figure if the fans are bad then so are the temps? Which sounds completely backwards to me - if the fans are suspect, temp monitoring is even more important. And while it's possible that one of the temp sensors is on the fan board itself, at least one must be on the main board. And it's more likely that both are.

And that still doesn't explain why temps work for a bit then stop. Perhaps the switch lets temp monitoring continue for a bit with 'bad' fans then after a period decides enough is enough. Which is also bizarre. Or it could just be a bug of course.

Anyway, it is what it is. At least it can be made to work again by using at least one 4-pin fan, or with a little extra work if it has to be 3-pin fans.
 

Corsaire

New Member
Feb 3, 2018
21
0
1
I could put my hand on a Base-T SFP I had tested before today and tested it my switch.

When running the Quanta firmware, no dice.
But with the brocade firmware works like a charm.

Not much info though.

Code:
SSH@switch10g>show media ethernet 13
Port     13: Type  : 1G M-TX(SFP)
 

Corsaire

New Member
Feb 3, 2018
21
0
1
Anyway, it is what it is. At least it can be made to work again by using at least one 4-pin fan, or with a little extra work if it has to be 3-pin fans.
Annnnnd it's back :)

I did have some 4-pin fans installed but not plugged on the 3rd position fan power plug.

Funny thing, I did the switch online and the temperature then showed 128 Celcius degrees... yeah right.

But after a reboot was back to more regular figures
Code:
SSH@switch10g>show chassis
Power supply 1 (NA - NA - Regular) present, status ok
Power supply 2 (NA - NA - Regular) present, status ok

Fan 1 ok, speed (auto): 1<->[[2]]<->3
Fan 2 ok, speed (auto): 1<->[[2]]<->3
Fan 3 ok, speed (auto): 1<->[[2]]<->3

Fan controlled temperature: 42.0 deg-C

Fan speed switching temperature thresholds:
                Speed 1: NM<----->30                            deg-C
                Speed 2:       25<----->65                      deg-C
                Speed 3:             60<----->90                deg-C (shutdown)


Exhaust Side Temperature Readings:
        Current temperature : 42.0 deg-C
        Warning level.......: 80.0 deg-C
        Shutdown level......: 90.0 deg-C
Intake Side Temperature Readings:
        Current temperature : 39.0 deg-C
Cacti shows 3 probes now
Management module: CPU temperature
Management module: MAC 1 temperature
Management module: MAC 2 temperature

I wish I knew about plugging the 4 pins fans in the 3rd plug before.
 

mixmansc

Member
Feb 15, 2016
46
26
18
Another semi-interesting bit. I mentioned in a prior post about the chips and also noted that the ram is a regular SODIMM. I also had mentioned that I also just bought an H3C switch that I crossed over to new HP firmware. On peeking under the hood of that switch I noticed it also used a basic SODIMM. So I looked again and both the LB6M and the H3C S5800 both use a single 512MB PC2-5300S (DDR 333) SODIMM. Being the dork I am I hopped on eBay and did a quick search and literally like the 4th item was this listing... 4GB Kit (4x 1GB) DDR2 PC2-5300S 2Rx8 Laptop SODIMM Memory | eBay I figure for five bucks and I get a couple of spares why not.

So five bucks and a week later the SODIMM's show up. Popped one in both the H3C and the Quanta and both seem perfectly happy and report that they now have 1024MB of ram instead of 512MB.

I seriously doubt it would ever make any difference whatsoever but why not... Maybe if there is some firmware bug that causes a memory leak I'll get a little more time before things crash or something. LOL

"turboiron#show memory
Total DRAM: 1073741824 bytes
Dynamic memory: 905527296 bytes total, 862785536 bytes free, 4% used"

It also reports 1024MB at boot. :)
 

schwartznet

New Member
Feb 7, 2018
8
1
3
40
Im tempted to get one of these and flash it. Has anyone figured out the LED's yet? Not a huge deal but would be nice to have.
 

fohdeesha

Kaini Industries
Nov 20, 2016
2,728
3,078
113
33
fohdeesha.com
Im tempted to get one of these and flash it. Has anyone figured out the LED's yet? Not a huge deal but would be nice to have.
Not possible to fix due to differences in architecture. One uses a CPLD to multiplex, the other uses native cpu I/O since it has half as many LEDs to control
 
  • Like
Reactions: TheBloke

mixmansc

Member
Feb 15, 2016
46
26
18
Yup - what might be more of an issue is what the other end is connected to. If just an HBA then most any will probably work, if using to uplink to another brand of switch then you might need to be sure the cable is supported by the other switch. I'm connecting mine to an HP switch and the HP bitches if its not an "HP" DAC cable. Fortunately it does still let me use it though. :)
 

Corsaire

New Member
Feb 3, 2018
21
0
1
I made more tests today toying with some optical SFP.

Here are the results I got
Code:
Port      7: Type  : 1G M-SX(SFP)
             Vendor: Brocade            Version:
             Part# : PL-XPL-VC-S13-19   Serial#: #######

Port     13: Type  : 1G M-SX(SFP)
             Vendor: Brocade            Version:
             Part# : QFBR-5767L         Serial#: #######


Port     13: Type  : 1G M-SX(SFP)
             Vendor: FINISAR CORP.      Version: X1
             Part# : FTRJ-8519-3        Serial#: ######
One thing however, I couldn't get any information from the command
Code:
# show optic thresholds PORT
Is it because those are some SFP, and not SFP+ ?