We actually can use UNII-1. How we get so damn many BSSIDs in UNII-1 I'm not sure. Very faint office Wifi, I suspect.
Take a look at an airview from Charlestown, MA to a rooftop in downtown Boston. You can see even UNII-2 is quite noisy. But UNII-1 is not bad. Interference is very different link-by-link, but in general, the more frequencies at our disposal, the better. Colin netBlazr <http://netblazr.com/> - free your broadband! The wireless telegraph is not difficult to understand. The ordinary telegraph is like a very long cat. You pull the tail in New York, and it meows in Los Angeles. The wireless is the same, only without the cat. On Mon, Nov 3, 2014 at 6:02 PM, Matt Hoppes <[email protected]> wrote: > Not to say this isn't an issue, but are you even able to use those > frequencies if there are that many APs around? > > On Nov 3, 2014, at 5:46 PM, Colin Zwiebel <[email protected]> wrote: > > This is correct. I titled the thread wrong, instead of "Major Bug in > v5.5.10" should have been "Major Bug in all v5.5." Any Ubnt STA that sees > more than 200 networks is unable to connect to anything. > > If we limit our scan list to only 28% of 5GHz frequencies (5160-5280) and > we still run into the *buffer to large(65535) for realocating* issue. > This means UNII-1 is very hard for us to use. > > I am also concerned that one day the cable goons will flip the switch, > double the number of BSSIDs they broadcast, we'll be over the 200 limit, > and half of our customers will disconnect. This BSSID issue is really a > problem for urban wisps. > > There was a proposed fix from the Linux community in 2009 > http://osdir.com/ml/linux-wireless/2009-02/msg00202.html > I don't know if it was adopted, but the proposed fix from 2009 was to > truncate the scan instead of choking and failing to scan at all. > > > *v5.6 to the rescue... but not yet* > AirOS version 5.6 (tested v5.6beta5) also has the issue. No fix quite yet. > > If the patch is simple, I would really like a v5.5 firmware that doesn't > have the *buffer to large(65535) *I would like to use UNII-1 as soon as > possible. It is true, scan list is a workaround, but we are going the wrong > way, making it more difficult to change frequencies instead of easier. > Because of urban interference issues, I would like change AP frequencies > nightly or so, measure statistics, and thus find low-interference channels > (we cannot scan in real-time and airview done at night does not reflect > daytime interference). As I am forced to make the scan list smaller and > smaller, I have to update scan-lists on all STAs before I can change AP > frequency, making it much harder to change freq nightly. > > > > http://community.ubnt.com/t5/Installation-Troubleshooting/v5-5-10-major-bug-no-SSIDs-in-scan-print-scanning-info-buffer/m-p/1077389/highlight/false#M72179 > > > > Colin > netBlazr <http://netblazr.com/> - free your broadband! > > > The wireless telegraph is not difficult to understand. The ordinary > telegraph is like a very long cat. You pull the tail in New York, and it > meows in Los Angeles. The wireless is the same, only without the cat. > > On Thu, Oct 30, 2014 at 2:49 PM, Matt Hardy <[email protected]> wrote: > >> To follow up on this thread, we found the issue. This wasn't really a new >> bug introduced in v5.5.10, but with the additional frequencies the customer >> was seeing > 100 new APs broadcasting in the UNII-1 band (probably lots of >> indoor APs). There's a limit of 200 APs the radio can pick up when doing a >> scan, if the radio is in an area with > 200 broadcasting SSIDs, it causes >> this issue (on any firmware) and device stops performing as expected. >> >> We'll be increasing the limit in v5.6, but in the meantime there are >> workarounds: scan list, different channel bandwidths, etc. >> >> Thanks, >> Matt >> >> On Tue, Oct 28, 2014 at 7:52 PM, Phil Curnutt <[email protected]> wrote: >> >>> Is there a Scan List set? Cut the number of channels it searches down >>> to a minimum. >>> >>> Phil >>> >>> On Tue, Oct 28, 2014 at 11:30 AM, Matt Hoppes < >>> [email protected]> wrote: >>> >>>> XW is different.... than XM. >>>> >>>> On 10/28/14, 12:35 PM, Josh Reynolds wrote: >>>> > So they say, yet when XW 5.5.10 was placed on XW Ti hardware, the gear >>>> > went nuts for a full 24 hours before we tried 5.5.6. When it was just >>>> as >>>> > bad, I went to 5.5.10rc3... same issues. 5.5.10rc2? Everything went >>>> back >>>> > to normal. It wasn't just one piece of hardware either, the same >>>> > symptoms were noticed on 4 different APs. >>>> > >>>> > They can say nothing changed all they want-- something changed, there >>>> > were 3rcs, a final, and now tons of bug reports on 5.5.10. >>>> > >>>> > Josh Reynolds, Chief Information Officer >>>> > SPITwSPOTS, www.spitwspots.com <http://www.spitwspots.com> >>>> > >>>> > On 10/28/2014 05:05 AM, Matt Hoppes wrote: >>>> >> You know there were no changes made to XM v5.5.10 other than adding >>>> new >>>> >> frequencies, right? >>>> >> >>>> >> On 10/27/14, 8:57 PM, Colin Zwiebel wrote: >>>> >>> We are having serious problems with v5.5.10 with 5GHz XM hardware >>>> and >>>> >>> have decided this firmware is not production ready. We using >>>> *v5.5.10 >>>> >>> only for links that require UNII-1* frequencies and avoiding in all >>>> >>> other situations. There is some sort of a bug with scanning for >>>> >>> available networks, so if your area has few networks, you might >>>> observe >>>> >>> different behavior. >>>> >>> >>>> >>> >>>> http://community.ubnt.com/t5/Installation-Troubleshooting/v5-5-10-major-bug-no-SSIDs-in-scan-print-scanning-info-buffer/m-p/1072997#U1072997 >>>> >>> >>>> >>> *AirOS v5.5.10 is not production ready!* >>>> >>> >>>> >>> Ubiquiti, it appears you have a major bug and need to revoke the >>>> >>> release, remove from website, until you get this fixed. >>>> >>> >>>> >>> >>>> >>> >>>> >>> We have now had 3 clients -- 2 Nanobridges and 1 Rocket M5 -- >>>> disconnect >>>> >>> from their AP with the same behavior: NO SSIDs appear in scan list. >>>> >>> Downgrade to v5.5.8, no changes to config, radio once against sees >>>> SSID >>>> >>> and connects perfectly. >>>> >>> >>>> >>> >>>> >>> >>>> >>> *Workaround: Frequency Scan List* >>>> >>> >>>> >>> I cannot fully confirm this, but it appears, v5.5.10 plus frequency >>>> scan >>>> >>> list has this problem less. It also might be that this pares down >>>> the >>>> >>> amount of data coming from the atheros driver (by not scanning all >>>> >>> frequencies), and scan list doesn't actually fix the issue, just >>>> reduces >>>> >>> the scan data to something that doesn't break the radio. >>>> >>> >>>> >>> >>>> >>> >>>> >>> >>>> >>> >>>> >>> *iwlist ath0 scan -- print_scanning_info: buffer too large(65535) >>>> for >>>> >>> realocating* >>>> >>> >>>> >>> In our most recent incident, I was able to SSH into the failed >>>> v5.5.10 STA >>>> >>> >>>> >>> XM.v5.5.10# iwlist ath0 scan >>>> >>> print_scanning_info: buffer too large(65535) for realocating >>>> >>> >>>> >>> XM.v5.5.10# wlanconfig ath0 list scan >>>> >>> ioctl[unknown???]: Argument list too long >>>> >>> wlanconfig: unable to get scan results >>>> >>> >>>> >>> >>>> >>> >>>> >>> We are running now with a frequency scan list (cannot downgrade this >>>> >>> radio) and both iwlist and wlanconfig commands are working properly. >>>> >>> >>>> >>> >>>> >>> >>>> >>> More details: >>>> >>> >>>> >>> APs: Problem happens with both Nanostation M5 APs and >>>> PowerBridge-M5 APs >>>> >>> >>>> >>> - Seen with AP on v5.5.10 >>>> >>> >>>> >>> - Seen with AP on v5.5.8 >>>> >>> >>>> >>> STAs: Problem observed with NB-22-M5 and Rocket-M5 >>>> >>> >>>> >>> DFS: Problem observed on both DFS and non-DFS frequencies (5805) >>>> >>> >>>> >>> >>>> >>> >>>> >>> Most recent incident, Rocket was working fine for a few days, >>>> suddenly >>>> >>> died. It rebooted in the process (only 2min uptime when I got to >>>> >>> it). Let sit 10min, rebooted, made no difference, gave buffer too >>>> larger >>>> >>> error no matter how many reboots. Only enabling scan list patched >>>> it. >>>> >>> >>>> >>> >>>> >>> >>>> >>> Colin >>>> >>> netBlazr <http://netblazr.com/> - free your broadband! >>>> >>> >>>> >>> >>>> >>> The wireless telegraph is not difficult to understand. The ordinary >>>> >>> telegraph is like a very long cat. You pull the tail in New York, >>>> and it >>>> >>> meows in Los Angeles. The wireless is the same, only without the >>>> cat. >>>> >>> >>>> >>> >>>> >>> _______________________________________________ >>>> >>> Ubnt_users mailing list >>>> >>> [email protected] >>>> >>> http://lists.wispa.org/mailman/listinfo/ubnt_users >>>> >>> >>>> >> _______________________________________________ >>>> >> Ubnt_users mailing list >>>> >> [email protected] >>>> >> http://lists.wispa.org/mailman/listinfo/ubnt_users >>>> > >>>> > >>>> > >>>> > _______________________________________________ >>>> > Ubnt_users mailing list >>>> > [email protected] >>>> > http://lists.wispa.org/mailman/listinfo/ubnt_users >>>> > >>>> _______________________________________________ >>>> Ubnt_users mailing list >>>> [email protected] >>>> http://lists.wispa.org/mailman/listinfo/ubnt_users >>>> >>> >>> >>> _______________________________________________ >>> Ubnt_users mailing list >>> [email protected] >>> http://lists.wispa.org/mailman/listinfo/ubnt_users >>> >>> >> >> _______________________________________________ >> Ubnt_users mailing list >> [email protected] >> http://lists.wispa.org/mailman/listinfo/ubnt_users >> >> > _______________________________________________ > Ubnt_users mailing list > [email protected] > http://lists.wispa.org/mailman/listinfo/ubnt_users > > > _______________________________________________ > Ubnt_users mailing list > [email protected] > http://lists.wispa.org/mailman/listinfo/ubnt_users > >
_______________________________________________ Ubnt_users mailing list [email protected] http://lists.wispa.org/mailman/listinfo/ubnt_users
