Hi Norman,
To me, 11ax APs shouldn't even be on the Enterprise market yet. I know that
doesn't touch your question, and we all have our own "you do what you gotta do"
realities.
Thanks for reading through that long post.
-Lee
Lee Badman | Network Architect (CWNE#200)
Information Technology Services
(NDD Group)
206 Machinery Hall
120 Smith Drive
Syracuse, New York 13244
t 315.443.3003 e lhbad...@syr.edu w its.syr.edu
SYRACUSE UNIVERSITY
syr.edu
-----Original Message-----
From: The EDUCAUSE Wireless Issues Community Group Listserv
<WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU> On Behalf Of Norman Elton
Sent: Friday, January 10, 2020 10:10 AM
To: WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU
Subject: Re: [WIRELESS-LAN] Who has transitioned away from Aruba, and why?
I agree with 100% of that. But here's a question ...
>> I absolutely will not sacrifice an otherwise sound WLAN by tweaking
>> configs or code upgradin for some small minority of poorly designed
>> or suddenly misbehaving clients that can be fixed from the client
>> side
What about Intel's AX driver bugs? I absolutely hate the idea of disabling AX
to support a few clients. But how many people are telling their helpdesk to
upgrade drivers on whatever BYOD laptop shows up?
What about a conference with 200 laptops that suddenly finds that half are
unsupported?
But, once it's disabled, will we ever re-enable AX? It's easy to say that we'll
disable it "short term", but we know those drivers won't magically update
themselves. We could be looking at crippling our wireless indefinitely :-/.
Our current AX test environment has it turned off on the 2.4 radio, so that at
least those users can connect someplace. Leave 5 GHz for those that can support
AX. I don't like the compromise, but the alternative ("hey we're trying out a
brand new wireless network that won't work for random people") is equally
unappetizing.
Sigh.
Norman Elton
William & Mary
On Fri, Jan 10, 2020 at 9:36 AM Lee H Badman
<000000db5b77bd95-dmarc-requ...@listserv.educause.edu> wrote:
>
> I know a lot of people are likely following along, so I’ll throw one more
> rant nugget out there (and this is not meant to distract from Ryan’s original
> question):
>
>
>
> Over the many years I’ve been doing this, I have found that MOST problems on
> a healthy, well-designed wireless network are absolutely client-related. Even
> on the likes of Active Directory managed PCs where the assumption is that
> Windows updates make everything fine. These updates don’t tend to touch WLAN
> adapter, BIOS, and chipset drivers which are often the root cause of wireless
> issues.
>
>
>
> Then there is the fallacy that the latest Intel/Broadcom driver is the
> “best”. Sometimes you have to use an older one on a specific model PC or NIC-
> especially where you are doing 802.1X. The whole effect is greatly magnified
> in the BYOD world that many of us live in with endless mainstream and not so
> mainstream client OS’s. Is it the WLAN vendor’s job to make up for all the
> goofy, ill-designed crap that’s out there? (Talking myself back from the
> ledge here, before I go off on the Wi-Fi Alliance). This situation sucks
> largely, and we’re stuck with it so we have to manage as best as we can.
>
>
>
> Then there are the optional features- for example, I’ve seen band
> steering make life tough for Windows PCs seemingly out of the blue.
> Except it wasn’t out of the blue- it was after Windows’ Patch Tuesday.
> In this case, disabling long-enabled band steering “fixed” the problem
> of users having wireless connectivity but not getting anywhere and
> losing massive amounts of pings. BTW… band-steering is not part of the
> 802.11 standard. Where does “fault” lie in this situation? Microsoft?
> The WLAN adapter/driver vendor? The WLAN vendor? Me? It’s messy as
> hell at times, given that “standards” are often a big fat lie when it
> comes to wireless in my opinion. Disagree? I’ll fight ya J
>
>
>
> So… my premise is that MOST of the time the clients are the issue. And for
> me, I absolutely will not sacrifice an otherwise sound WLAN by tweaking
> configs or code upgrading for some small minority of poorly designed or
> suddenly misbehaving clients that can be fixed from the client side, and I
> don’t hold any WLAN vendor responsible for fixing the endless list of issues
> in the client space.
>
>
>
> But when infrastructure code deficiencies DO hit, and all of the optional
> features have been disabled and all of the client devices have been proven to
> be as healthy as they can be first, it’s the worst of the worst situations
> for those of us who run big networks because it’s truly out of our hands.
> While I don’t expect Cisco or Aruba or whoever to make up for client
> shortcomings or to jump through hoops so some unholy bizarre feature can be
> implemented (vendors do TOO MUCH of this, in my opinion), I do expect the
> vendors to absolutely keep their own houses in order and to understand that
> in big university settings STABILITY IS EVERYTHING.
>
>
>
> If code is bad, tell us. Tell everyone, proactively. Get it the hell off of
> the website so no one else downloads it. Don’t leave us in “we need to gather
> data” status- that’s why vendors have million dollar test facilities (and
> I’ve seen many of them)- gather your own data and just get us back on the
> rails. If code is considered “bleeding edge”, be honest about that with big
> red warning labels on the UI and the download links. If HW is defective- same
> thing. Recall it. Proactively. If HW is “bleeding edge” be brutally honest.
> Customers should not be part of the QA process or have to play code roulette
> to find what is “safe”. Any vendor who dares charge for a “bug scrub” before
> recommending a good code version in this Age of Crappy Code should be ashamed
> of themselves, BTW.
>
>
>
> And finally. any vendor or VAR who can cavalierly say “well the customer
> bought bleeding edge stuff, what do they expect” has lost touch with what
> customer service means. If it’s that fragile, it shouldn’t be on the market,
> period. Silly Vally needs to slow it down. It ain’t Agile if it sucks.
>
>
>
> Sorry for the rant.
>
>
>
> Lee Badman | Network Architect (CWNE#200)
>
> Information Technology Services
> (NDD Group)
> 206 Machinery Hall
> 120 Smith Drive
> Syracuse, New York 13244
>
> t 315.443.3003 e lhbad...@syr.edu w its.syr.edu
>
> SYRACUSE UNIVERSITY
> syr.edu
>
>
>
> From: The EDUCAUSE Wireless Issues Community Group Listserv
> <WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU> On Behalf Of Michael Davis
> Sent: Friday, January 10, 2020 7:31 AM
> To: WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU
> Subject: Re: [WIRELESS-LAN] Who has transitioned away from Aruba, and why?
>
>
>
> FWIW, some of the most bizarre issues I've ran into with Aruba APs have been
> related to:
> - MTUs on the path
> - Reassembly of packets
> - Out of order fragments
> - LLDP
> - tx, beacon, basic radio rates
>
> Some things to look into if the 5GHz radio drop can be
> deterministically recreated and tested, but I know that's usually half the
> battle..
>
>
> On 1/9/20 3:34 PM, Turner, Ryan H wrote:
>
> We are on 8.5.0.3 for the ITS cluster. We were going to upgrade to 8.0.0.5,
> but we had a disaster in one of our data centers just before the holidays.
> Power was tripped for a 13,000 sq foot data center. For some reason, APs
> associated to the controller in this building did not fail over to the other
> site. We are going to be testing this scenario again next week by yanking
> the power to confirm if we’ve hit yet another bug, or if this was a one-off.
>
>
>
> Ryan
>
>
>
>
>
> From: The EDUCAUSE Wireless Issues Community Group Listserv
> <WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU> On Behalf Of Steve Fletty
> Sent: Thursday, January 9, 2020 1:20 PM
> To: WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU
> Subject: Re: [WIRELESS-LAN] Who has transitioned away from Aruba, and why?
>
>
>
> What version of 8.5?
>
>
>
> We saw some issues in our lab prior to 8.5.0.4. We have a mix of 335s and
> 535s.
>
>
>
> On Thu, Jan 9, 2020 at 10:15 AM Turner, Ryan H <rhtur...@email.unc.edu> wrote:
>
> All:
>
>
>
> We’ve been an Aruba shop for a very long time and have around 10,000 access
> points. While every relationship with vendors have their ups and downs, my
> frustration with the Aruba is finally peaking to the point that I am
> considering making the enormous move to choose a different vendor. The
> biggest reason is with the 8.X code train, and bugs that we just don’t
> consider appropriate to use in production. It has been one thing after the
> other, and my extremely talented and qualified Network Architect (Keith
> Miller) might as well be on the Aruba payroll as much work as he has been
> doing for them to solve bugs. Just when we think we have one fixed, another
> one crops up.
>
>
>
> The big one as of late is with 515s running 8.5 code train. We have them
> deployed in one of our IT buildings. Periodically, people that are connected
> to these APs in the 5G band will stop working. To the user, they are
> browsing a site, then it becomes unresponsive. If they are on their phone,
> they will disconnect from wifi and everything works fine on cell. Nothing
> makes an 802.11 network look worse than switching to cell and seeing a
> problem resolve. Normally, if the users disconnect then reconnect, their
> problems will go ahead (but I think they end up connecting in the 2.4G band).
> We’ve been working on this problem with them for months. It always seems
> as though we have to prove there is a real issue. I’m fed up with it. We
> are a sophisticated shop. If we have a problem, 9 times out of 10 when we
> bring it to the vendor, it is a real problem. I’m extra frustrated that due
> to issues we’ve seen in ResNet on the 8.3X train that we don’t want to
> abandon our 6 train on main campus. To Aruba’s credit, we purchased around
> 1,000 515s last year (I think around February). When they could not get good
> code to support them on, Aruba bought back half of them. I asked for them to
> buy back half because I thought for sure with the 315s that we would have
> instead, the issues would be fixed by the time the 315s ran out. Not looking
> to be the case.
>
>
>
> So, with that rant over, we are seriously considering looking to move away
> from Aruba (unless they get their act together really soon). There are other
> bugs I’m not even mentioning here. For those of you that made the switch to
> another vendor, I would be curious how long the honeymoon lasted, what were
> your motivators, and were you happy with the overall results? Of course,
> this is a great opportunity to plug your vendor. As I see it, we have 3
> choices…. Something from Cisco (we had Cisco long ago and dumped them for
> bugs), something from Extreme (we are a huge Extreme shop so this makes
> sense), something from Juniper (Mist).
>
>
>
>
>
> **********
> Replies to EDUCAUSE Community Group emails are sent to the entire
> community list. If you want to reply only to the person who sent the
> message, copy and paste their email address and forward the email
> reply. Additional participation and subscription information can be
> found at https://www.educause.edu/community
>
> **********
> Replies to EDUCAUSE Community Group emails are sent to the entire
> community list. If you want to reply only to the person who sent the
> message, copy and paste their email address and forward the email
> reply. Additional participation and subscription information can be
> found at https://www.educause.edu/community
**********
Replies to EDUCAUSE Community Group emails are sent to the entire community
list. If you want to reply only to the person who sent the message, copy and
paste their email address and forward the email reply. Additional participation
and subscription information can be found at https://www.educause.edu/community
**********
Replies to EDUCAUSE Community Group emails are sent to the entire community
list. If you want to reply only to the person who sent the message, copy and
paste their email address and forward the email reply. Additional participation
and subscription information can be found at https://www.educause.edu/community
