Aruba MAS series SFP+ & POE+ switches sub-$100

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

ViciousXUSMC

Active Member
Nov 27, 2016
264
140
43
41
2 fans in and 2 out will make for excellent airflow; it's referred to as 'balanced draft'. Just because the internal pressure is held near a zero delta to the outside doesn't mean there's no air flow...

Power generating station scenario: petroleum coke fired, top-supported watertube steam generators producing superheated steam which is expanded through steam turbines to generate power. Process control software holds the furnace pressure to -0.5", while twin, steam-turbine driven forced air fans pressurize the unit at the burner deck and a slip stream gets directed to secondary air fans then to ball mills where it mixes with/ carries pulverized coke to the burners. At the exhaust plenum, a single induced-draft fans take suction from the furnace and delivers it to the electrostatic precipitators for ash removal then to the stack. All fans have adjustable inlet vanes to vary airflow.

The furnace is held to a slightly negative pressure so that air infiltration is inward, not outward. Internal restrictions to air flow include a feedwater pre-heater (economizer) and both primary and secondary superheaters (steam must be heated beyond its saturation temperature to prevent condensation on blades as the steam is expanded through turbines. Condensation causes pitting and can lead to deposits on the blades resulting in unbalanced conditions and power loss. Blade washing is a thing.)

This scenario is common to all fired boilers, whether they have just forced air fans or both forced & induced air systems, due to a condition known as 'complete combustion'. Target combustion is a slight excess O2 and zero CO in the exhaust: too much excess O2 and you're wasting fuel by heating too much air and sending it up the stack; too little O2 and the fuel is partially un-burnt and creating CO as a product of incomplete combustion. CO is still capable of being burned, and in large enough % is explosive, and can lead to a 'furnace pressure excursion'. Redundant instrumentation at the stack continuously measures exhaust O2, CO2, CO, NOx, SO2 and particulate matter.

Zero pressure differential does not necessarily indicate zero air flow.

I worked 16 years in this environment; 10 years on the floor and 6 in the control room. 450MW generating station attached to a heavy oil upgrader.
Pretty sure that entire thing was summed up in my first sentence:
No it does not, but it also depends a lot on the conditions.
I do not have a background like that, but I do have a background in building computers, speakers, etc and I know from experience what has worked for me and what has not.

I mean if your turning this switch into a steam turbine, by all means give it a go.
And above all else, like I said to the last person. Try it and report back.
 

Dawg10

Associate
Dec 24, 2016
220
114
43
And above all else, like I said to the last person. Try it and report back.
I learned early on that P.Eng's are smarter than me. If a product is designed to work in one manner, any other manner is to be avoided unless re-certified by another P.Eng. Reversing or otherwise modifying switch airflow means you're running the product in a manner not tested nor approved/ certified by the manufacturer nor any associated licensing body. Yes, modifying switch fans is as minor an issue as you can get, but the underlying comment is the same: you don't know what unintended/ negative impact you're making on the product.

In a switch with low internal resistance to air flow, 2 fans in and 2 fans out is very nearly the equivalent of just 2 fans in, which is likely why it doesn't work for you.
 

myrison

Member
Jan 26, 2011
61
41
18
I'm definitely not smart enough to comment on the previous discussion, but did want to update everyone that I ordered a couple of the fans below to see how they far in the switches. I think that my switch came with a couple of the fans that may have been bad/aging (i.e. bad bearings or something similar that caused more of a 'whine' than they should have while running).

Evercool 40x40x20 4-pin fan

I'll let everyone know once I have a chance to give them a run.
 

ViciousXUSMC

Active Member
Nov 27, 2016
264
140
43
41
The stock fans are two different type in my switch.
I lost the pictures I took of both but the one I found is the Delta Electronics FFB0412SHN
13000RPM
24CFM
54.5DBA

This gives you a good place to start on replacement fans.

For reference the Noctua NF A4x20 FLX I used are.
5000RPM
5.53CFM
14.9DBA

I think good replacements would be things like:
Sunon 40mm (40x40x20mm) 12V DC Fan Model KDE1204PKV3.MS.AR.GN RPM Sensor
5200RPM
6.8CFM
18DBA

So a bit louder, a bit stronger, and about half the price of the Noctua fans.

The other thing is the stock fans are 28mm depth and there is a delicate balance between speed, noise, and size.
The smaller the fan the less CFM and more speed/noise it needs to compensate. The easy to find fans were 10mm and 20mm so giving up 8mm of depth for the Noctua meant just by default I lost some potential power.

On the coolerguys site they do have 28mm fans, so those would be the ideal replacements.
Something like this: Evercool 40x40x28mm 12V PWM Fan with Connector-EC4028HH12BP

Should pretty much end up being a direct replacement in size/capabilities but a bit slower and a bit quieter.
 
  • Like
Reactions: vdmann

myrison

Member
Jan 26, 2011
61
41
18
Thanks Vicious. FYI the ones I ordered are the smaller 20mm depth ones.

Info about these varies a bit on different sites... but this is what I believe they offer:

40x40x20, model EC4020SH12BP:
4-pin
Noise: 14.7- 32 dBA
Speed: 3000 +/- 25% ~ 9000 +/- 10%RPM
Airflow 9 CFM max

The 40x40x28mm one Vicious posted (model EC4028HH12BP):
4-pin
Noise: 16.1- 46 dBA
Speed: 3000 +/- 25% ~ 12000 +/- 10%RPM
Airflow 19 CFM max

For those that aren't familiar, the decibel approximately doubles in perceived loudness every 10dB, so the 28mm fan is approximately 2x as loud and 2x as powerful as the 20mm version.

For the 20mm fans I chose, compared to the Noctuas, they are capable of being as quiet as the Noctuas at their lowest RPMs but will be about 3x as loud. (and 30 dB online is described as nearby whispering... if that's what they produce I can live with that) :)

Ultimately the proof will be in the listening once they're here, but on paper at least they seem to fall nicely in the middle (which was pure luck for me as I didn't realize there was a 28mm version of the fan until Vicious posted it).
 

ViciousXUSMC

Active Member
Nov 27, 2016
264
140
43
41
One more for you guys. Covers creating new user accounts, vlans, jumbo frames, switching-profiles, etc.


If I do any more it will be advanced with access control lists or something as I think this covers most of the basics.
However if you have ideas, let me know!
 

klui

Well-Known Member
Feb 3, 2019
824
453
63
If I do any more it will be advanced with access control lists or something as I think this covers most of the basics.
However if you have ideas, let me know!
That was a good video--thanks. The overall concept of assigning a profile to a interfaces is not too difficult.

One thing I've never done is ACLs and if you use them at work (I don't deal with networking, only from a periphery level) I would be interested getting your take about ACLs on a single subnet. For instance, different organizations using a single /22. I saw a document from Cisco that describes in certain cases you need two (or maybe more, I gather) ACLs due to how the mask is set up and that's the thing that was difficult to grasp.
 

cheezehead

Active Member
Sep 23, 2012
723
175
43
Midwest, US
That was a good video--thanks. The overall concept of assigning a profile to a interfaces is not too difficult.

One thing I've never done is ACLs and if you use them at work (I don't deal with networking, only from a periphery level) I would be interested getting your take about ACLs on a single subnet. For instance, different organizations using a single /22. I saw a document from Cisco that describes in certain cases you need two (or maybe more, I gather) ACLs due to how the mask is set up and that's the thing that was difficult to grasp.
So in a larger subnet say no one needs access to the security cameras but now one machine does...ACL's can be used to allow the one machine through to security camera subnet. At home this is pretty limited, but for work they really extend this functionality through security roles/policies via clearpass so that when user-x logs into anywhere they would receive the same security profile. If another person logs into the same workstation a minute later, they would get a different security profile.
 

ske4za

Member
Feb 4, 2019
80
45
18
ViciousXUSMC, thanks for the video again. Do you know how I can do UDP broadcast forwarding on this switch? On my Procurve 2910AL, I can go into
Code:
int vlan <vlan-id>; ip forward-protocol udp <port> <ip address>
if I want to forward a UDP port across VLANs (my example is forwarding UDP 65001 from my wifi VLAN to my main VLAN for to be able to stream from an HDHomeRun). Based on my research, it looks like you have to
Code:
int vlan <vlan-id>; ip directed-broadcast
to enable directed-broadcast, create an ACL profile that matches the UDP protocol, port, and IP address, and then apply it to the VLAN interface. I'll give it a shot when I have some time this week, but I think this is the way I need to go.
 

ViciousXUSMC

Active Member
Nov 27, 2016
264
140
43
41
I should look into that myself and see. I have a digital mixer on a different vlan than my computer, and I always have to hand type my static IP to connect to it.

All I found is this, but I do not know these commands are accurate: ArubaOS 7.3
 

ske4za

Member
Feb 4, 2019
80
45
18
The current documentation can be found here (Aruba 7.4.X, if you are on the latest firmware). I skimmed through it, looks like you can create an alias through netservice with the specific UDP port, and then apply it to an ACL on an interface (VLAN or otherwise). I haven't tested it yet.
 

Emulsifide

Active Member
Dec 1, 2014
212
93
28
Just to confirm DAC compatibility, I'm currently experimenting with some HP DL360p v2 servers and these switches. I'm able to confirm connectivity between the servers-->switches and switch-->switch with the following knock-off J9283B equivilant:

https://www.amazon.com/gp/product/B078YSWDS6/

EDIT: To be clear, the servers have HP 530FLR NICs installed in them (part number 647581-B21/684210-B21).
 
Last edited:

ske4za

Member
Feb 4, 2019
80
45
18
Just to confirm DAC compatibility, I'm currently experimenting with some HP DL360p v2 servers and these switches. I'm able to confirm connectivity between the servers-->switches and switch-->switch with the following knock-off J9283B equivilant:

https://www.amazon.com/gp/product/B078YSWDS6/
Thanks for that. I bought a DAC that wouldn't work from the S2500 (not Aruba certified either) to an X520 card. I did some research and bought a Molex one off eBay which is another option and a bit cheaper. 74752-1301 Molex 10GB SFP+ 3M Passive Cable SFP-H10Gb-Cu3M | eBay

Code:
GE0/1/3
-------
Vendor Name                                : Molex Inc.
Vendor Serial Number                       : 435732358
Vendor Part Number                         : 74752-1301
Aruba Certified                            : YES
Cable Type                                 : 10GBASE-DAC-P
Connector Type                             : Copper Pigtail
Wave Length                                : 0 nm
Cable Length                               : 3m
And from the other side:

Code:
ifconfig -v ix0
ix0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=e400bb<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,JUMBO_MTU,VLAN_HWCSUM,VLAN_HWTSO,RXCSUM_IPV6,TXCSUM_IPV6>
        ether 0c:c4:7a:ba:b3:55
        hwaddr 0c:c4:7a:ba:b3:55
        inet6 fe80::ec4:7aff:feba:b355%ix0 prefixlen 64 scopeid 0x1
        inet 10.0.10.2 netmask 0xfffffffc broadcast 10.0.10.3
        nd6 options=21<PERFORMNUD,AUTO_LINKLOCAL>
        media: Ethernet autoselect (10Gbase-Twinax <full-duplex,rxpause,txpause>)
        status: active
        plugged: SFP/SFP+/SFP28 1X Copper Passive (Copper pigtail)
        vendor: Molex Inc. PN: 74752-1301 SN: 435732358 DATE: 2014-12-23
 
  • Like
Reactions: Arbies

Emulsifide

Active Member
Dec 1, 2014
212
93
28
Bummer. I tend to not go below $25.00 for a DAC knowing that most of the crap ones only work with low-end equipment. I've modified my above comment to state which NICs are in the servers.
 

sfair

New Member
Jan 4, 2018
13
3
3
38
I ordered one earlier when the listed was a bit newer.

Vicious, Thanks for all the info.

My switch came with a newish firmware. It also had the Aruba rack ears so that's a bonus! The rack ears put the switch further back than Cisco switches. The front of the switch is pretty close to flush with the front of the rack ears similar to most patch panels.

I upgraded to the latest firmware from the HPE site via the web interface. I also removed the two stacking ports via CLI as instructed.

The fans are loud on startup but otherwise not too bad when throttled down to low speed. The Aruba switch replaces an older Cisco 2960 and 3560 in my rack. The blower style Cisco 2960 was a screamer and required a fan mod to quiet it down. I like to run Noctua fans wherever possible but I'm not sure if I want to spend the $60 to quiet it down further yet.
 

myrison

Member
Jan 26, 2011
61
41
18
Hey guys - I mentioned earlier I'd post back after trying the fans mentioned below.

In summary, I'd recommend these fans over the Noctuas, particularly if you have the PoE version of the switch. My PoE switch was running at a temp of 99 Celcius on DEP-0 with 4 Noctua fans, and after replacing 2 of the Noctuas with the fans below, my average temps have dropped by 5-7 degrees average across all measurements. The switch also remains very very quiet. MUCH quieter than with stock fans.

If anyone has their stock fans still in, and has a PoE-48 switch, can you post your temps and note how many of the PoE ports you're using so I can see what temps a 'stock' switch runs at?

Ideally post the temps after it's been running a few hours under a real load as they definitely start much cooler and heat up over the first few hours of use with a real load running through them.

Using 25 ports plugged in, 5 of them PoE, my temps with the modified 2x Evercool fans, 2x Noctua fans are:


System Temperature
: CPU Temp : 63 C

: DPE-0 Temp : 87 C

: DPE-1 Temp : 69 C

: DDR Temp : 56 C

: PSU Temp0 : 53 C

: PSU Temp1 : 60 C

: Exhaust Temp : 58 C

: Mid Temp : 63 C

: Ambient Temp : 51 C

: Console Card Temp : 34 C




Info about these varies a bit on different sites... but this is what I believe they offer:

40x40x20, model EC4020SH12BP:
4-pin
Noise: 14.7- 32 dBA
Speed: 3000 +/- 25% ~ 9000 +/- 10%RPM
Airflow 9 CFM max

For the 20mm fans I chose, compared to the Noctuas, they are capable of being as quiet as the Noctuas at their lowest RPMs but will be about 3x as loud. (and 30 dB online is described as nearby whispering... if that's what they produce I can live with that) :)

Ultimately the proof will be in the listening once they're here, but on paper at least they seem to fall nicely in the middle (which was pure luck for me as I didn't realize there was a 28mm version of the fan until Vicious posted it).
 
Last edited:

ViciousXUSMC

Active Member
Nov 27, 2016
264
140
43
41
I will probably fire up my PoE switch soon. It's only the 24P model though it may have a bit of impact but probably not much since they have the same PoE budget.

Also I would be sure to ask what the ambient temp is, that makes a HUGE difference in temps.
 

Dawg10

Associate
Dec 24, 2016
220
114
43
@myrison your temperature numbers cannot be right. Ambient temperature of 51C...? Exhaust temp of 58C while the CPU and mid temps are each 63C...? These are not correct.

I played with the sequence a couple times: starting with the ambient being the coolest @ 34C and inverting the order results in:


: CPU Temp : 63 C

: DPE-0 Temp : 58 C

: DPE-1 Temp : 60 C

: DDR Temp : 53 C

: PSU Temp0 : 56 C

: PSU Temp1 : 69 C

: Exhaust Temp : 87 C

: Mid Temp : 63 C

: Ambient Temp : 34 C

: Console Card Temp : 51 C

This gives you believable numbers: ambient (coolest) @ 34C and exhaust (hottest) @ 87C.

Do you have access to a IR heat gun? This may be the only way to confirm.
 

sic0048

Active Member
Dec 24, 2018
114
95
28
I'm running the 48 port POE switch, but I only have 19 1000/100/10 devices plugged into it and of those, only 8 are POE devices (and all of those are pretty low power - currently drawing anywhere from 0.9 to 4.1v ). Stock fan configuration.......

: CPU Temp : 50 C

: DPE-0 Temp : 68 C

: DPE-1 Temp : 60 C

: DDR Temp : 43 C

: PSU Temp0 : 38 C

: PSU Temp1 : 43 C

: Exhaust Temp : 44 C

: Mid Temp : 50 C

: Ambient Temp : 42 C

: Console Card Temp : 30 C