Page 1 of 1

Nstreme test package

Posted: Fri Oct 17, 2008 3:58 pm
by sergejs
We have optimized the Nstreme code, that performs much better for point to multipoint environment. It is necessary to use new Nstreme package on Access Point router, client routers can use old version, but we recommend to install new package for clients too.

Packages for download:

mipsbe package;
mipsle package;
ppc package;
x86 package.

For more information and any problems with test package, please contact MikroTik support (support@mikrotik.com).

Re: Nstreme test package

Posted: Fri Oct 17, 2008 11:40 pm
by nz_monkey
Hi sergejs

Can you explain what has changed in Nstreme ?

Re: Nstreme test package

Posted: Sat Oct 18, 2008 12:19 am
by edward08
Should I install this upgrade in a point-to-point scenario?

Re: Nstreme test package

Posted: Mon Oct 20, 2008 11:08 am
by uldis
We have improved the polling mechanism for nstreme.
Please let us know if that solves the problem with the lot of wireless nstreme clients (more than 30).
About point-to-point, you can also install it on the point-to-point links, but the most benefits are for point-to-multipoint.

Re: Nstreme test package

Posted: Mon Oct 20, 2008 10:41 pm
by shoffmann
I have put this package on ~20 various AP's and it has tremendously improved AP's and customer performace on mipsbe and mipsle chipsets only. On PowerPC units with Nstreme on it has absolutely destroyed the throughput, latency and jitter.

Has anyone else seen this?

Steve
Pocketinet Communications

Re: Nstreme test package

Posted: Tue Oct 21, 2008 11:00 am
by uldis
I have put this package on ~20 various AP's and it has tremendously improved AP's and customer performace on mipsbe and mipsle chipsets only. On PowerPC units with Nstreme on it has absolutely destroyed the throughput, latency and jitter.

Has anyone else seen this?

Steve
Pocketinet Communications
Please make the support output file from your PowerPC AP when you see worse throughput with the new test package. Send it to support@mikrotik.com

Re: Nstreme test package

Posted: Wed Oct 22, 2008 12:20 pm
by uldis
we have updated these test packages, please download them again. There was a bug (kernel panic) when creating support output file with regular wireless mode (not nstreme).

Re: Nstreme test package

Posted: Thu Oct 23, 2008 9:24 pm
by infowest
Any clues on what has changed? Fragmentation? Any config changes we need to plan on to optimize this, or is it all automagic?

thanks!

Randy

Re: Nstreme test package

Posted: Thu Oct 23, 2008 10:59 pm
by tjohnson
Also, you "recommend" upgrading the clients, but it was not necessary. Is there a benefit to upgrading the clients? Some more info about the changes would be very helpful.

Re: Nstreme test package

Posted: Thu Oct 23, 2008 11:36 pm
by believewireless
When trying this on an RB532 we get the printout at the end of the message. If we can log in fast enough and disable the test package, everything is fine. This happens on two different mipsle boards.


MikroTik RouterOS 3.15 (c) 1999-2008 http://www.mikrotik.com/

dec/31/2001 20:00:05 system,error,critical System rebooted because of kernel
fai
l
ure
dec/31/2001 20:00:05 system,error,critical router was rebooted without proper
sh
u
tdown (cause 1)
dec/31/2001 20:00:08 system,error,critical System rebooted because of kernel
fai
l
ure
dec/31/2001 20:00:08 system,error,critical router was rebooted without proper
sh
u

Re: Nstreme test package

Posted: Thu Oct 23, 2008 11:39 pm
by tjohnson
I have installed on several RB532 boards with no problems.

Re: Nstreme test package

Posted: Thu Oct 23, 2008 11:58 pm
by Mplsguy
Polling algorithm has been adjusted to be more suitable for large numbers of clients and for most common traffic patterns (bursty traffic with rather low average rate). As polling is controlled entirely by AP and protocols are backwards compatible, only AP is required to be upgraded.

There are two reasons why it is recommended to upgrade clients as well - first, there are some improvements for clients to increase efficency (but that should not produce notable improvement, therefore it is not required), second, all development happens with new test package on both AP and clients, therefore only this combination can be considered best optimized. Although reports on test results when using older clients are welcome as of now, you can expect that support will ask you to upgrade some time later, when that will matter.

As of now, test package does not require any configuration changes/adjustments - consider it all "automagic". New wireless configuration settings that have been added to wireless-test are not pertinent in this case as they do not affect nstreme operations.

Fragmentation of packets has been available for nstreme before (controlled with framer-* settings). These improvements do not change or affect that.

There are some other improvements in development that should further help jitter issues over PtMP nstreme links, so you can expect new test packages soon. Feedback in form of supout files and detailed problem descriptions are welcome.

Re: Nstreme test package

Posted: Fri Oct 24, 2008 1:13 am
by rpingar
we tested the new package on two misple 532 and it is fine.

One AP have 44 clients and another 30.
In both cases the new package performed better them the regualr one. The latency when down from 40-60ms in the first cast o 6-10ms.
BUT

but the jitter is still here, in fact it does several ping to 6-10ms and then goes to 120ms for some seconds then it is back to normale, or on packet at 6ms and the following at 120ms....

So i think the poll mechanisme has provided a very good performance but MT should still work on jitter.

Then letting us to use queue in the way to reporduce WMM should close the gap to a professional system.

Regards
Ros

Re: Nstreme test package

Posted: Fri Oct 24, 2008 12:29 pm
by uldis
We have updated the wireless-test package. Please download them again, upgrade your routers and report back how it is working.

Re: Nstreme test package

Posted: Fri Oct 24, 2008 2:08 pm
by rpingar
Tested release of today 24/10
very very promising.

Tested only on mipsle.
Higher CPU usage and this is a nice thing for wireless customers.....

very few spike in latency, and bettere average performance on latency.

Good work.

Ros

Re: Nstreme test package

Posted: Sat Oct 25, 2008 2:09 am
by Alessio Garavano
We have updated the wireless-test package. Please download them again, upgrade your routers and report back how it is working.
When was the rebuild of the wireless-test package? i downloaded theses in 10/20/2008...

Re: Nstreme test package

Posted: Sat Oct 25, 2008 9:45 am
by rpingar
they were updateted yesterday.

We tested them on three wireless router and only on one we had a little isse about changing on the fly the Hardware Retries, it crashed the router.
Disabling the interface changing the parameter and then re enabling the wireless worked around the problem.

The performance of last build is fantastic.
GREAT WORK MT.

Regards
Ros

Re: Nstreme test package

Posted: Sat Oct 25, 2008 8:19 pm
by shoffmann
I'm testing this now on a ppc board with 25 clients. It is working quite well.

Previous version did not work well at all on the ppc boards.

Thanks for the great work.

Steve
Pocketinet

Re: Nstreme test package

Posted: Mon Oct 27, 2008 10:01 am
by uldis
they were updateted yesterday.

We tested them on three wireless router and only on one we had a little isse about changing on the fly the Hardware Retries, it crashed the router.
Disabling the interface changing the parameter and then re enabling the wireless worked around the problem.

The performance of last build is fantastic.
GREAT WORK MT.

Regards
Ros
did it create an autosupout.rif file? if yes then please email it to support@mikrotik.com

Re: Nstreme test package

Posted: Mon Oct 27, 2008 10:49 pm
by BrianHiggins
Put simply, AMAZING IMPROVMENTS!!

we've tested on 3 of our heaviest loaded / worst performing PtMP APs, 2 of them are using SR9 radio's and the other has a XR5. the SR9 APs (one 532 and one 333) have shown improvments from average latency of 100-140ms (jitter ranging from 20ms to 300ms) down to now 15-20ms average latency and jitter ranging from 7 to 40ms

Additionally radio uptimes on the SR9 links seem to be noticably better with much less bouncing (momentary disconnect / near instant reconnect) of the connections.

the 5.8ghz AP showd even less jitter, and average latency of about 8ms to most clients! (improved from 30 to 40ms previously)

will be testing on a few more APs over the next few days assuming these 3 don't show any problems.

all around this seems to be a wonderfull update, and I strongly encourage people to test this out in the field and report back their experiences.

Re: Nstreme test package

Posted: Tue Oct 28, 2008 11:21 am
by Muqatil
I've done some field test:

On RB532 with 2 wlan:
CPU 100% all the time
Good ping and jittering on all 30 users on each wlan

On RB411 with 1 wlan used as PtP
CPU 100% when there's a lot of traffic
Kernel failure and resets.

On RB333 with 3 wlan
CPU always 100%
every wlan couldn't support more than around 15 users (they usually had 25 users each). Users keep disconnecting

Except for the RB532, i had to remove the wireless-test package from the machines to prevent crashes.
I am sorry i totally forgot to create supout from those.

Re: Nstreme test package

Posted: Tue Oct 28, 2008 5:34 pm
by shoffmann
I will have to back up a bit from my previous post. I tested this firmware on several of my AP's for two days. It seemed to work quite well. I have now rolled 10 more AP's to it and have a mess. Customers flooding my support lines with close to no bandwidth. Continuous re registering of SU's (some every 15 sec.). Even the ones that were working fine have seemed to crash.

I will send a support.rif

Steve
Pocketinet

Re: Nstreme test package

Posted: Wed Oct 29, 2008 4:09 pm
by uldis
please download the new packages - we have updated them.
New packages should contain fix for crash and optimize the disconnection problem.
Report back with the results.

Re: Nstreme test package

Posted: Wed Oct 29, 2008 6:14 pm
by Muqatil
now mipsle crashes and ppc doesn't...
sent supout btw

Re: Nstreme test package

Posted: Thu Oct 30, 2008 6:23 am
by BrianHiggins
I have seen few oddities with the new package (but still overall love the improvments!)

we have a tower that has two 900mhz 120 deg h-pol sectors, one is powerd by a SR9, the other by a XR9, both are in 5mhz channels, on opposite ends of the 900mhz band.

If we install the new test package on either of the APs, all radio links on the other AP not running the test package immeaditly become unstable and have trouble even connecting. removing the test package from the SR9 AP on 907 mhz fixes the connection problems on the XR9 AP on 922 mhz (both are in 5mhz channel mode).

Haven't been able to get it to work stabily on a 333 yet, haven't tried the new new package, will give that a try tomorrow.

EDIT (forgot to include the 2nd odd issue):

When a SR9 CPE is idle, it (or all CPEs on the AP) will momentarily disconnect / reconnect after a few hours, but if you launch a ping test to a CPE (even if it was previously only staying connected for a few minutes at a time), it will not disconnect at all for at least 24+ hours (have tested 2 seperate CPEs over a 24 hour period each).

It seems there may be some minor bug in the new package that causes radio's with little or no traffic across them to drop, and it seems that it happens to nearly all of the CPEs at the same exact time, except the one with a constant ping test running against

Re: Nstreme test package

Posted: Thu Oct 30, 2008 8:52 am
by 0ldman
So far, so good. 5GHz backbone, 2.4GHz and 900MHz AP, all's well. Jitter is almost gone, ping has improved greatly all the way across the board, just like before I enabled Nstreme.

I can ping google at 11ms (was 35 on average) and my farthest CPE, two hops, 14 miles at < 5ms.

Re: Nstreme test package

Posted: Thu Oct 30, 2008 11:45 pm
by postmanerk
First, GREAT improvements on the NStreme package. Lower ping times (100ms to 20ms or so with 50 clients). Major improvements. Jitter is still not totally consistent, but better.

I am seeing low up times (connectivity from clients). I have been on 3.10 and uptimes of days (30-80 days continuous). Now on 3.15 with the test package, I see maybe an hr then they all re-connect. Only a few stay connected at 1 hr or more.

Anyone else see low client connection times? Should I sent a support file?

Eric

Re: Nstreme test package

Posted: Fri Oct 31, 2008 1:26 am
by Alessio Garavano
With wireless-test package the remote nodes of my mesh network are disconnecting continuosly....

I have an RB600 with 3 5Ghz cards(2xR52H-1xXR5) like the backbone of the net, the disconnect problem is continuosly with the older test-package and with the new builded 29/10.

Disabling the wireless-test and enabling the normal 3.15 wireless package all is stable and running perfectly

Re: Nstreme test package

Posted: Fri Oct 31, 2008 2:39 am
by 0ldman
My system is more stable. 5MHz channels, if it makes a difference.

Re: Nstreme test package

Posted: Fri Oct 31, 2008 11:03 am
by uldis
First, GREAT improvements on the NStreme package. Lower ping times (100ms to 20ms or so with 50 clients). Major improvements. Jitter is still not totally consistent, but better.

I am seeing low up times (connectivity from clients). I have been on 3.10 and uptimes of days (30-80 days continuous). Now on 3.15 with the test package, I see maybe an hr then they all re-connect. Only a few stay connected at 1 hr or more.

Anyone else see low client connection times? Should I sent a support file?

Eric
Yes, send the support output file to support@mikrotik.com when you see this problem with the wireless-test package. Make sure that you use the latest wireless-test package that is posted here in the forum.

Re: Nstreme test package

Posted: Fri Oct 31, 2008 11:29 am
by chvdr
what about polling interval at 5GHz-turbo. what should it be.

Re: Nstreme test package

Posted: Fri Oct 31, 2008 2:52 pm
by mietus
quick question..
where is package for x86? :>

Re: Nstreme test package

Posted: Fri Oct 31, 2008 4:24 pm
by BrianHiggins
I am seeing low up times (connectivity from clients). I have been on 3.10 and uptimes of days (30-80 days continuous). Now on 3.15 with the test package, I see maybe an hr then they all re-connect. Only a few stay connected at 1 hr or more.

Anyone else see low client connection times?

Eric - Check out my post here...
http://forum.mikrotik.com/viewtopic.php?f=7&t=27708

While testing the new package I came across something else, It was more a general NStream topic so I didn't think it should go in this thread, but it is definitly worth noting. Pay specific attention to the section about the SR9 CPE that wouldn't stay connected for more then 5-10 seconds while running the new test package...

Uldis: Can you confirm anything in regards to the Max Station Count and the new test package? It looks to me like the old package determines part of the polling algorithm on whatever this value is set to as it seems to have a dramatic effect on latency, does the new test package dynamically alter this portion of the algorithm to always match the number of CPE's connected or somthing similar?

Re: Nstreme test package

Posted: Sat Nov 01, 2008 5:56 pm
by peavys
On a troublesome 900mhz sector in a noisy environment with 12 clients it made a big difference. With or w/o nstreme in the past I had very bad latency. I don't see any other improvements, but the ping times quieted dramatically.

I've never considered nstreme a realistic option for pToMpt links in the past but I'll have to re-consider that. Unfortunately I've started deploying nano stations though, so it still won't be an option on most APs.

Clients still drop off/on just like they did before, and speeds are still the same. CPU consumption runs around 25%, but I did not note what it was before. I tried it with max station count set to 20 and 2000, and that part did not seem to make any difference.

Corky
nstreme.PNG

Re: Nstreme test package

Posted: Mon Nov 03, 2008 12:04 pm
by uldis
please send the support output files from your routers where you have installed the wireless-test package to support@mikrotik.com so we could check how well it is working.

Re: Nstreme test package

Posted: Mon Nov 03, 2008 6:42 pm
by BrianHiggins
please send the support output files from your routers where you have installed the wireless-test package to support@mikrotik.com so we could check how well it is working.

do you mean everyone who has it installed, or are you just referring to peavys?

Re: Nstreme test package

Posted: Mon Nov 03, 2008 7:26 pm
by tomswenson
I installed this package and turned on nstreme (I'm pretty new to mikrotik wireless, but not mikrotiks) for the first time and had nothing but trouble with disconnections, so I had to go off of nstreme. I didn't get any support files as I was in a panic to fix things. I used it on two ap's with about a half dozen users each.

Tom

Re: Nstreme test package

Posted: Tue Nov 04, 2008 3:38 am
by 0ldman
Stability was down. One CPE that stays connected with a weak signal was virtually unusable with the test package.

Re: Nstreme test package

Posted: Tue Nov 04, 2008 1:54 pm
by uldis
please send the support output files from your routers where you have installed the wireless-test package to support@mikrotik.com so we could check how well it is working.

do you mean everyone who has it installed, or are you just referring to peavys?
I ment everyone :)

Re: Nstreme test package

Posted: Tue Nov 04, 2008 7:09 pm
by Alessio Garavano
With 3.15 and Wireless-Test my 25 remote nodes in 5Ghz are disconnecting continuosly. Each node is serving at 40-50 CPEs

Uninstalling wireless-test and with the 3.15 regular wireless package, the disconnection problem was solved, but the performance of the net is very poor at moments, i have 5 or 6 times on the day when the network is very slow 3000-4000ms pings to each remote node by about 10 or 15 minutes. Then all work great, but is a big problem.

Now i downgrade all my RBs to 3.13 and the network is again running excellent...

I think the problem is not only the wireless-test, the problem is the 3.15

Regards,

Re: Nstreme test package

Posted: Wed Nov 05, 2008 3:44 pm
by uldis
If you upgrade to v3.16, then for the nstreme improvements you should enable the wireless-test package.

Re: Nstreme test package

Posted: Wed Nov 05, 2008 5:21 pm
by shoffmann
Just upgraded one of my sectors to the above package. I don't see any change from the 3.15 test package yet. While the upload and download are consistent on some others continue to associate and disassociate with virtually non-existent upload. Latency remains < 100ms on most during heavy traffic…that is except for the ones that are rolling.

support file sent

Steve
Pocketinet Communications

Re: Nstreme test package

Posted: Wed Nov 05, 2008 5:59 pm
by shoffmann
Further data on 3.16 wireless test package installed this morning

I find that the cpe's that are disassociating are fine and stay associated with download but have virtually no upload and associate/disassociate continually with upload. this problem is on a sector with 17 cpe's with the mipsbe board. Another sector with the same number of cpe's and the powerpc bd. works Great.

Steve
Pocketinet Communications

Re: Nstreme test package

Posted: Thu Nov 06, 2008 6:45 pm
by ropebih
Hi every1,
I have installed wireless-test package on my 3 Point to Point links and with that I solved problem with unstable ping (1 - 30ms). Ping is now max 3ms and I'm thrill but now I have new problem. Link is broken every 1 or 2 hours.
03:47:11 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: disconnected, extensive data loss 
03:47:16 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: connected, wants WDS 
05:00:16 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: disconnected, extensive data loss 
05:00:21 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: connected, wants WDS 
05:59:45 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: disconnected, extensive data loss 
05:59:50 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: connected, wants WDS 
06:07:13 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: disconnected, extensive data loss 
06:07:19 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: connected, wants WDS 
07:01:51 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: disconnected, extensive data loss 
07:01:56 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: connected, wants WDS 
08:44:13 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: disconnected, extensive data loss 
08:44:18 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: connected, wants WDS 
08:57:41 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: disconnected, extensive data loss 
08:57:46 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: connected, wants WDS 
11:02:59 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: disconnected, extensive data loss 
11:03:04 wireless,info 00:0B:6B:86:xx:xx@GLOBAL STOZER 2 - Bugojno: connected, wants WDS

Re: Nstreme test package

Posted: Thu Nov 06, 2008 6:54 pm
by 0ldman
I'm having the same upload problem with an RB600 with Crossroads clients.

Re: Nstreme test package

Posted: Sat Nov 08, 2008 10:37 am
by ropebih
on x86 with wireless-test:
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=40 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=50 ms
10.10.60.9 64 byte ping: ttl=64 time=40 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
10.10.60.9 64 byte ping: ttl=64 time=30 ms
with standard wireless package:
10.10.60.9 64 byte ping: ttl=64 time=1 ms
10.10.60.9 64 byte ping: ttl=64 time=1 ms
10.10.60.9 64 byte ping: ttl=64 time=1 ms
10.10.60.9 64 byte ping: ttl=64 time=1 ms
10.10.60.9 64 byte ping: ttl=64 time=1 ms
10.10.60.9 64 byte ping: ttl=64 time=1 ms
10.10.60.9 64 byte ping: ttl=64 time<1 ms
10.10.60.9 64 byte ping: ttl=64 time<1 ms
10.10.60.9 64 byte ping: ttl=64 time=10 ms
10.10.60.9 64 byte ping: ttl=64 time=10 ms
10.10.60.9 64 byte ping: ttl=64 time=1 ms
10.10.60.9 64 byte ping: ttl=64 time=10 ms
10.10.60.9 64 byte ping: ttl=64 time=1 ms
10.10.60.9 64 byte ping: ttl=64 time=1 ms

Re: Nstreme test package

Posted: Mon Nov 10, 2008 12:37 pm
by uldis
ropebih, please make support output file when you use regular wireless package and the also one when you use wireless-test package.

Re: Nstreme test package

Posted: Mon Nov 10, 2008 3:41 pm
by ste
Uldis,

will there be the possibility to do a mixed Nstream/Non Nstream Segment.
At the moment we use Nstream only for Backhaul links.
So when this new version is leaving test-state we want to switch to Nstream.
But we cannot migrate all clients at once so for a limited time we need to
run both modes at once.
Another (better) solution would be when clients (all MT) have a autodetection
and switch Nstream on/off as needed to connect to the ap. In this scenario
we would update all clients and then turn Nstream on at the AP.

Regards,
Stefan

Re: Nstreme test package

Posted: Mon Nov 10, 2008 4:42 pm
by uldis
currently it isn't possible.You should have nstreme or regular wireless link.

Re: Nstreme test package

Posted: Tue Nov 11, 2008 2:48 am
by ropebih
Done, I sent both supout files. I allso noticed that on wireless-test package clock is very fast. He's showing 10-15 secs in 1 "real" second

Re: Nstreme test package

Posted: Tue Nov 11, 2008 1:33 pm
by Muqatil
RB532 with 3.16 wireless-test:
Impressive improvement on connection... It's stable and excellent ping.. 100%CPU though...
RB333 with 3.16 wireless-test:
Good improvement on connections.. but it's too much instable... Clients keep disconnecting "Extensive data loss". When i rollback on wireless normal package, those Clients are stable. CPU 100%


RB532 got 2x AR5006X 802.11abg NIC (rev: 1)
RB333 got 3x AR5212 802.11abg NIC (rev: 1)

Re: Nstreme test package

Posted: Tue Nov 11, 2008 2:22 pm
by uldis
RB532 with 3.16 wireless-test:
Impressive improvement on connection... It's stable and excellent ping.. 100%CPU though...
RB333 with 3.16 wireless-test:
Good improvement on connections.. but it's too much instable... Clients keep disconnecting "Extensive data loss". When i rollback on wireless normal package, those Clients are stable. CPU 100%


RB532 got 2x AR5006X 802.11abg NIC (rev: 1)
RB333 got 3x AR5212 802.11abg NIC (rev: 1)
Please provide us with the support output files from your RB333 where you see instability. We need one support output file with wireles-test where you see those disconnects and one with the regular wireless package that is running fine.
Also did you see this issue with v3.15 wireless-test package?

Re: Nstreme test package

Posted: Tue Nov 11, 2008 5:10 pm
by Muqatil
RB532 with 3.16 wireless-test:
Impressive improvement on connection... It's stable and excellent ping.. 100%CPU though...
RB333 with 3.16 wireless-test:
Good improvement on connections.. but it's too much instable... Clients keep disconnecting "Extensive data loss". When i rollback on wireless normal package, those Clients are stable. CPU 100%


RB532 got 2x AR5006X 802.11abg NIC (rev: 1)
RB333 got 3x AR5212 802.11abg NIC (rev: 1)
Please provide us with the support output files from your RB333 where you see instability. We need one support output file with wireles-test where you see those disconnects and one with the regular wireless package that is running fine.
Also did you see this issue with v3.15 wireless-test package?
Sent.
If I recall correctly, there was a 3.15 package that worked pretty fine on RB333 but i updated it (and cancelt previous) that was similar to 3.16 one.

Re: Nstreme test package

Posted: Wed Nov 19, 2008 4:36 pm
by tjohnson
Any more status on this? Any new updates?

Re: Nstreme test package

Posted: Wed Nov 19, 2008 7:13 pm
by Wavelan
Bug in 3.16 with wireless.test (X86)

bug in P4 hw :
1. system clock is 10 times faster than normal

Bug in wireless (XR5 & XR5)

1. Compression doesn't work, cannot connect in station-wds and bridge mode
if compression is enabled.

2. Erratic troughput. Goes from 33 Mbps, to 8 mbps, back to 30 mbps.
If we're change back to 3.14, troughput is stable around 33 mbps.

Thank's
//Robert

Re: Nstreme test package

Posted: Thu Nov 27, 2008 11:51 am
by acim
Done, I sent both supout files. I allso noticed that on wireless-test package clock is very fast. He's showing 10-15 secs in 1 "real" second
Same in my case, x86 Core2 Duo. Clock is much faster, although ping seems to be smaller and performances better. Can you resolve this with clock, please? I will let you know is something else goes bad.

Re: Nstreme test package

Posted: Thu Nov 27, 2008 12:29 pm
by acim
Because of broken clock, other time related matters are also broken, for example bandwidth test, interface traffic meter and so. Please fix this ASAP, otherwise new package seems to work really better, specially pings.

Re: Nstreme test package

Posted: Thu Nov 27, 2008 12:31 pm
by normis
new build seems to have fixed the clock, everyone with clock problems, please report to support, you will get the new version

Re: Nstreme test package

Posted: Thu Nov 27, 2008 12:42 pm
by acim
normis, how to get the new build? I've already reported to support, maybe I should just wait a little bit :)

Re: Nstreme test package

Posted: Thu Nov 27, 2008 12:47 pm
by normis
ask about new build that fixes clock issue :)

Re: Nstreme test package

Posted: Thu Nov 27, 2008 5:17 pm
by ejansson
I have seen the massive and fast disconnects too. I have rolled back to the regular 3.16 package and things were much better, I can send the .rif of both, but I need to flip it back to the test package. How much time do I need to let it run to capture enough data, I would like to minimize the customer impact.

Erik

Re: Nstreme test package

Posted: Mon Dec 01, 2008 1:58 pm
by uldis
We need support output file from regular v3.16 package and v3.16 wireless-test package.
Few minutes would be fine - so we could see in the logs what mac address disconnect.

Re: Nstreme test package

Posted: Wed Dec 03, 2008 12:51 pm
by cybernet
Hi All, How many clients can connect to AP-Bridge with NSTREME test package? I use ROS 3.16 and wireless test package.

Re: Nstreme test package

Posted: Fri Dec 05, 2008 4:54 pm
by Muqatil
Did you change something in wireless-test-3.17-ppc.npk from 3.16?

Re: Nstreme test package

Posted: Sun Dec 07, 2008 9:31 pm
by 0ldman
3.17 test seems more stable and ping times have dropped.

Several customers on my 2.4GHz AP have 1 day + uptime, which I've only seen on rare occasion. Looks good so far.

Re: Nstreme test package

Posted: Sun Dec 07, 2008 10:51 pm
by acim
I would say that wireless-test 3.17 suffers from the same clock problems as previous versions. Otherwise, it might be better, but it should not be installed on SMP routers. Mikrotik, please fix this ASAP, we want to test wireless-test :D

Re: Nstreme test package

Posted: Mon Dec 08, 2008 4:43 pm
by uldis
in v3.17 the wireless-test package you can use the hw-retries setting for the nstreme also. This setting will work just like in the regular mode. For bad links - the bigger the hw-retries the higher possibility to get bigger latency, the smaller the hw-retries the bigger possibility to get disconnects with "extensive data loss". You can tune this setting to get optimal results in each case.
About x86 clock problem, we hope we will be able to fix it in v3.18.

Re: Nstreme test package

Posted: Mon Dec 08, 2008 6:27 pm
by tomswenson
I have a location with 3 3650 sectors. I have been watching this thread and testing the nstreme as I would like to use it if possible. Today, I took the 3.17 and wireless test and put it on one of the sectors with about 8 cpes. It all seemed to be fine. No one was reconnecting over and over like before. However, I found out that one of the other sectors at this location was now dropping and reconnecting some clients. I tried turning on the CSMA, but it didn't help. So, I switched off the nstreme for the sector I was testing it on and the other ap stabilized and is just fine now. The other two sectors are running 3.13.

Is the Nstreme not going to work in a multiple sector location or do I have to have it turned on in all ap's?

Tom

Re: Nstreme test package

Posted: Tue Dec 16, 2008 11:39 pm
by digicomtech
Uldis,

will there be the possibility to do a mixed Nstream/Non Nstream Segment.
At the moment we use Nstream only for Backhaul links.
So when this new version is leaving test-state we want to switch to Nstream.
But we cannot migrate all clients at once so for a limited time we need to
run both modes at once.
Another (better) solution would be when clients (all MT) have a autodetection
and switch Nstream on/off as needed to connect to the ap. In this scenario
we would update all clients and then turn Nstream on at the AP.

Regards,
Stefan
use a script to automagically switch on the client side:
/system script 
add name="ChangeBand" \
    source=":local delais 5s\r\n:local liste 2.4ghz-b/g,2ghz-5mhz,2ghz-10mhz;\r\n \
    \r\n:foreach band in=\$liste \
    do={\r\n:if (![/interface wireless get wlan1 running]) do={\r\n/interface \
    wireless set wlan1 band=\$band\r\n:delay \$delais\r\n}\r\n}\r\n\r\n:if \
    (![/interface wireless get wlan1 running]) \\\r\ndo={\r\n    :if \
    [/interface wireless nstreme get wlan1 enable-nstreme] \\\r\n    do={\r\n  \
          /interface wireless nstreme set wlan1 enable-nstreme=no\r\n    } \
    else={\r\n        /interface wireless nstreme set wlan1 \
    enable-nstreme=yes\r\n    }\r\n}"
/system scheduler 
add comment="" disabled=no interval=20s name="schedcheckWifi" \
    on-event=ChangeBand start-time=startup
Uldis, i got a many disconnect on x86 with wireless-test-3.17 package with this log message:
disconnected, too many poll timeouts

What does it mean ?

Re: Nstreme test package

Posted: Wed Dec 17, 2008 7:45 am
by 0ldman
I'm picking up pieces of your script here and there, but the short of it, what does it do?

Checks for 20/10/5, nstreme on or off, etc, and I assume it stops once its connected?

Re: Nstreme test package

Posted: Wed Dec 17, 2008 5:27 pm
by Muqatil
I think i found a pattern of my issues with wireless-test
in Radios with a 10° Antennas, i have no problems at all.. and it makes an HUGE improvement onf stability and throughput. No matter which type of radios of RBs i use..
in Radios with 30° or 60° or 90° antennas, i'm experiencing disconnections of users with good signal and good ccq, and troublesome users are unable to connect at all (CPEs that already have a poor signal but stay connected with not big issues without wireless-test)

I sent a lot of Supout but they couldn't find the problem. (I didn't tell them which type of antenna i used)

Can anyone confirm this?

Re: Nstreme test package

Posted: Fri Dec 19, 2008 3:29 pm
by acim
wireless-test 3.18 is working fine on Core 2 Duo for the time being.

Re: Nstreme test package

Posted: Mon Dec 22, 2008 8:34 pm
by acim
I still haven't noticed any discrepancy, everything is working fine, great pings 8)

Thank you MT :D

Re: Nstreme test package

Posted: Wed Dec 24, 2008 2:40 am
by exe
Did you include this improvements in:
http://www.mikrotik.com/download/wirele ... mipsbe.npk ?

Re: Nstreme test package

Posted: Wed Dec 24, 2008 8:29 am
by acim
No, I am talking about wireless-test 3.18 which is not public available yet. It is working fine with multicore routers, no clock problems described in this thread.

Re: Nstreme test package

Posted: Fri Dec 26, 2008 12:19 pm
by Hansanet
Hi ,

Where can I get v3.18 packages , I woud like to test it.

Thanks

Re: Nstreme test package

Posted: Sat Dec 27, 2008 4:02 am
by exe
My Analyze of 3.17 wireless package

I tested 3.17 wireless test package. It seems that there are problems: with clock, bandwith is very unstable, maximum 4M upload at Dual Nstreme Links, download is okay. It also has bad influence to other parts of network which is mikrotik based, so it become erratic.

********************************

Then, I tested 3.17 regular wireless package. I do not know if you implement here latest improvements, but news is very good! Great Work Guys!

Good:
1. Lower Ping: maximum 25ms on single Nstreme PTP, and most time it is 1ms...
2. No "eratic" influence to other parts on network, like before, especially those which have dual nstreme. Now you can use station wds in combination with single nstreme without a problem.
3. More stable bandwith on dual nstreme links.
4. Achieved 50Mbit/sec max at single nstreme ptp at low ping, which is very good.
5. No "sent deauth" problem as you can use station wds without risking to damage network.

Very Good Work!!!!!

Bad:
I will send this opinion to support.

Re: Nstreme test package

Posted: Mon Jan 05, 2009 3:42 am
by CarulloS
I posted this already in the wireless parent of this thread before I saw this... Just wanted to put it here too seemed like the right place...

Either my understanding needs adjustment or the option doesn't work as expected...

Using wireless test 3.17 on ap and clients i check and uncheck the default forward option on the ap and it makes no difference. Either way broadcast and normal ip traffic pass to all clients of the ap. From my understanding unchecking the DEFAULT FORWARD option would prevent traffic from passing from one client to another.

The AP is normal AP mode with default set to add all clients to wds bridge1 as well as the ethernet and wlan1 interfaces added... Make the ap act like any other trango etc ap just passing traffic from wireless to ethernet interface.

The clients are all station wds setup same as ap - one bridge eith ap added ad wds dynamically and eth1 and wlan1 added as well... again just passing traffic though to router behind the radio.

Thanks,
Scott

Re: Nstreme test package

Posted: Mon Jan 05, 2009 4:22 am
by InoX
I posted this already in the wireless parent of this thread before I saw this... Just wanted to put it here too seemed like the right place...

Either my understanding needs adjustment or the option doesn't work as expected...

Using wireless test 3.17 on ap and clients i check and uncheck the default forward option on the ap and it makes no difference. Either way broadcast and normal ip traffic pass to all clients of the ap. From my understanding unchecking the DEFAULT FORWARD option would prevent traffic from passing from one client to another.

The AP is normal AP mode with default set to add all clients to wds bridge1 as well as the ethernet and wlan1 interfaces added... Make the ap act like any other trango etc ap just passing traffic from wireless to ethernet interface.

The clients are all station wds setup same as ap - one bridge eith ap added ad wds dynamically and eth1 and wlan1 added as well... again just passing traffic though to router behind the radio.

Thanks,
Scott
I think the client will not see each other in Network Neighbour...but at IP level data will pass, ofcourse.

Re: Nstreme test package

Posted: Tue Jan 06, 2009 12:07 am
by CarulloS
They see each other with ip traffic and broadcasts (ie neighborhood list)

Re: Nstreme test package

Posted: Tue Jan 06, 2009 2:31 am
by Gerard
Did you see my reply in the other thread?

http://forum.mikrotik.com/viewtopic.php?f=7&t=28857

-Gerard

Re: Nstreme test package

Posted: Sun Jan 11, 2009 2:41 am
by ropebih
Wireless test package 3.18 have same clock problems as previous versions!!!

Re: Nstreme test package

Posted: Mon Jan 12, 2009 3:00 pm
by uldis
Wireless test package 3.18 have same clock problems as previous versions!!!
Please make the support output file and send it to support@mikrotik.com
Ayone else still have this problem or it is gone in v3.18?

Re: Nstreme test package

Posted: Mon Jan 12, 2009 4:19 pm
by Muqatil
Is it possible to have a pseudo-changelog of wireless test package? :o

Re: Nstreme test package

Posted: Wed Jan 14, 2009 5:47 am
by CarulloS
In 3-18-test mipsbe hit reset configuration on wireless and winbox closes and it does not reset. Confirmed multiple instances of this behavior.

Scott

Re: Nstreme test package

Posted: Wed Jan 14, 2009 11:35 am
by uldis
In 3-18-test mipsbe hit reset configuration on wireless and winbox closes and it does not reset. Confirmed multiple instances of this behavior.

Scott
I could not reproduce this problem.
Did the router create a autosupout.rif file, if yes then send it to support@mikrotik.com

Re: Nstreme test package

Posted: Thu Jan 15, 2009 12:43 pm
by ste
Hi,

anyone using 3.18 wireless-test with nstream in production?

I do some tests in lab at the moment. Looks promising but I see strange
effects so I do not dare to use it in production.

I'm testing RB500 - (Ethernet) - RB411 -(WLAN)- RB600 -(WLAN)- RB433 - (Ethernet) - RB500.
Btest UDP from RB500 to RB500 gives 60MBit using 5GHz Turbo. Replacing the RB600 by a RB500 (399MHz)
gives 40MBit. RB600/500 is in AP/Bridgemode on both links.

Then I tried to use 5GHz without Turbo and I see strange effects. First I changed one of the links to
non turbo mode on both sides of the link. The connection do not establish again. I have to disable/enable
the wlan at the RB600 (Master). Then it ran for a while. I tried to limit data rate at the RB600. Same
thing. I've to disable/enable at the Master side. The client has 2 errors: "can not connect unspecified(1)"
and "lost connection not polled for too long". Also I see changing speeds of the bandwith test. Sometimes
one of the 2 links goes down and reestablishes.
Wireless conditions are ideal. I use Omnis (MARS) at the RB600 and change antenna gain until all signals are
between 40 and 60.

So if I would have a production system with this version I would not dare to change anything as the link
might not come up again.

Any other experiences to share?

Stefan

Re: Nstreme test package

Posted: Sun Jan 18, 2009 3:12 pm
by zaltec
I have "lost connection not polled for too long" error too when using 3.18 wireless-test if there are interferences, using regular package is working quite well, even multiple links on the same frequency. So it seems wireless-test is not very good if you don't have a clean spectrum, which in our area is impossible (usually we see at least 15-25 radios active on 5.5-5.7 ghz band).

Regards,
Giorgio

Re: Nstreme test package

Posted: Sun Jan 18, 2009 6:02 pm
by ste
I have "lost connection not polled for too long" error too when using 3.18 wireless-test if there are interferences, using regular package is working quite well, even multiple links on the same frequency. So it seems wireless-test is not very good if you don't have a clean spectrum, which in our area is impossible (usually we see at least 15-25 radios active on 5.5-5.7 ghz band).

Regards,
Giorgio
I've used it in house. Frequency usage shows no other equipment. If interference it's self-interference.
I used 5,5 and 5,7 for the 2 connections. All cards are R52.

Regards,
Stefan

Re: Nstreme test package

Posted: Tue Feb 03, 2009 3:29 am
by ilius168
Are these updates included in the latest ROS Version (3.20) ? Or do we still need to update the wireless-test package?

Re: Nstreme test package

Posted: Tue Feb 03, 2009 4:23 pm
by uldis
you still need to use wireless-test package if you want those new features.

Re: Nstreme test package

Posted: Sat Feb 07, 2009 1:10 pm
by ropebih
I still have a problem with clock issue. ROS 3.20 x86 celeron 1,7 GHz.

Re: Nstreme test package

Posted: Sun Feb 08, 2009 3:19 am
by ilius168
Just want to report failure to run nstreme using wireless test package on RB 600.

Works fine and great on RB 411A, RB 411AH, RB 433, RB 433AH and RB 333.

Supout.rif file sent to support.

Thanks and great job guys!

Re: Nstreme test package

Posted: Wed Feb 11, 2009 9:03 pm
by samcarney
how can I roll back to my 3.10 package? 3.20 isnt working welll for me. I downloaded 3.13 and would like to try that one.

Re: Nstreme test package

Posted: Fri Feb 13, 2009 5:38 pm
by uldis
I still have a problem with clock issue. ROS 3.20 x86 celeron 1,7 GHz.
Try to enable the multi-cpu option to yes.

Re: Nstreme test package

Posted: Fri Feb 13, 2009 5:39 pm
by uldis
how can I roll back to my 3.10 package? 3.20 isnt working welll for me. I downloaded 3.13 and would like to try that one.
what exactly isn't working?
upload the packages to FTP and exacure the downgrade command in the system packages menu.

Re: Nstreme test package

Posted: Thu Mar 05, 2009 1:08 am
by infowest
I've got this working on a point to point link, nice and solid, no problems (433ah's). Getting ready to roll another pair with 411ah's and R5H's. Not the best test bed, but glad it is stable.

Any updates on what's new in this, what we can expect performance wise, etc? Especially interested in how many nstreme clients we can expect to connect before running into polling overload.

thanks!

Randy

Re: Nstreme test package

Posted: Thu Mar 26, 2009 8:48 am
by proggams2
do we still need the wireless test package for 3.22 ?? or we use the normal wireless package (in 3.22) ?? what i mean is the new nstream technology.

Re: Nstreme test package

Posted: Thu Mar 26, 2009 12:55 pm
by uldis
do we still need the wireless test package for 3.22 ?? or we use the normal wireless package (in 3.22) ?? what i mean is the new nstream technology.
You still need to use wireless-test package if you want to get those benefits that are discussed in theis topic.
Only if you install v4.0beta2, the wireless-test package is now regular wireless package.

Re: Nstreme test package

Posted: Fri Mar 27, 2009 4:14 pm
by angboontiong
Actually which version of the test package is reliable now???
And, how i going to have the test package? just a concern is, our link is in live network and easily running about 40Mbps, will it take risk if i load the say firmware??

Re: Nstreme test package

Posted: Wed Jun 03, 2009 10:30 pm
by iam8up
What is the minimum ROS version you need to apply the wireless-test package to see the ptmp improvements?

Re: Nstreme test package

Posted: Thu Jun 04, 2009 3:30 am
by InoX
Actually which version of the test package is reliable now???
And, how i going to have the test package? just a concern is, our link is in live network and easily running about 40Mbps, will it take risk if i load the say firmware??
Why do you think it's called a "test package"? Because it's a test package.
Stupid questions, get stupid answers.

Re: Nstreme test package

Posted: Thu Jun 04, 2009 7:47 pm
by infowest
I may be wrong, but I don't think there has been much change if any after 3.16.

We're using it everywhere now on 3.20-3.24 without problems, ptp and ptmp.

Re: Nstreme test package

Posted: Thu Jun 04, 2009 7:58 pm
by rpingar
i agree,
wireless-test should go to wireless legacy!

regards
ros

Re: Nstreme test package

Posted: Tue Jun 23, 2009 10:33 am
by titius
where can I get this wireless test packege?

Re: Nstreme test package

Posted: Wed Jul 01, 2009 4:48 pm
by chapex

Re: Nstreme test package

Posted: Wed Aug 05, 2009 10:34 am
by kabelfrei
zaltec wrote:
I have "lost connection not polled for too long" error too when using 3.18 wireless-test if there are interferences, using regular package is working quite well, even multiple links on the same frequency. So it seems wireless-test is not very good if you don't have a clean spectrum, which in our area is impossible (usually we see at least 15-25 radios active on 5.5-5.7 ghz band).

Regards,
Giorgio


I've used it in house. Frequency usage shows no other equipment. If interference it's self-interference.
I used 5,5 and 5,7 for the 2 connections. All cards are R52.

Regards,
Stefan
I have installed ROS V3.27 on a PtMP Installation (411A as AP; 1xRB532 Client; 1x RB333 Client; 1x RB433 Client) with wireless test package and got the same problems. Without nstreme enabled it worked. Also with nstreme enabled BUT Polling disabeld it worked. When nstreme with Polling is enabled, the 3 clients disconnects all the time (every ~10-30sec) with
AP-Log:
wireless,info 00:0B:XX:XX:XX:XX@wlan_xr5_ptp_sektor: disconnected, extensive data loss
Client-Log:
wireless,info 00:XX:XX:XX:XX:XX@wlan_r52_wka: lost connection, not polled for too long
We have also more then one Link on this location, but the next Links are on 5600 MHz and 5810 MHz.

Are there some additional config parameters needed on the AP side (perodic calibratio is on default (=disabeld), ANI is off)

thanks,

kabelfrei

Re: Nstreme test package

Posted: Wed Aug 12, 2009 9:18 pm
by kabelfrei
Nobody has an idea? I look at http://wiki.mikrotik.com/wiki/Wireless_Debug_Logs but i can´t find any entry for my Problem
wireless,info 00:XX:XX:XX:XX:XX@wlan_r52_wka: lost connection, not polled for too long
.

thanks,

kabelfrei

Re: Nstreme test package

Posted: Wed Sep 23, 2009 2:26 am
by acim
There is still the same problem with wireless-test package from version 3.30 on some single CPU platforms.

Older Compaq desktop with 730MHz Intel III CPU runs clock 10 times faster.

Newer machine with Core2 Duo CPU runs clock normally.

Is it possible to fix this for all platforms, please?

Re: Nstreme test package

Posted: Wed Sep 23, 2009 7:25 pm
by proggams2
Does nstreme work with the XR2 in routeros ?? OR any wireless brand other than mikrotik ??

Re: Nstreme test package

Posted: Wed Sep 23, 2009 8:17 pm
by CarulloS
Yes but your post has nothing to do with this thread.

Re: Nstreme test package

Posted: Fri Mar 05, 2010 4:49 pm
by xezen
have problem in version 4.6


error comes up

lost connection,not polled for to long

using xr5 nstream and and rb433ah

Re: Nstreme test package

Posted: Sat Mar 06, 2010 8:51 am
by 0ldman
3.30 is still *slightly* more reliable on noisy or weak links than the test. Most of my network runs wireless-test. I have 4 customers that are stable but low signal strength, with test they are not stable anymore. With 4.x, I can barely get into the equipment to change them back to 3.30.

Re: Nstreme test package

Posted: Mon Apr 12, 2010 11:48 am
by rmichael
[quote="0ldman"]3.30 is still *slightly* more reliable on noisy or weak links than the test. quote]

Is that with Nstreme and pooling ON?

Re: Nstreme test package

Posted: Mon Apr 12, 2010 8:20 pm
by 0ldman
With or without nstreme, polling, etc.

Re: Nstreme test package

Posted: Mon Apr 12, 2010 10:20 pm
by rmichael
With or without nstreme, polling, etc.
For comletness: Point to Point or Multipoint? I'm asking as v4.5 nstreme does not work for me (WDS/PtM; XR9) and I'm looking for best alternative. At the moment I'm considering v4.5 at end points and v3.3 at the base...

thanks,

rmichael

Re: Nstreme test package

Posted: Tue Apr 13, 2010 1:09 am
by Muqatil
I'm deploying wireless-test since the first release and i love it,
it tolerates less the bad links (that's why you say the normal package is more reliable), but those links are working bad anyway. If you raise the hw-retries, you get the same result as the normal package.
About 4.x, i had some problems at the beginning. i found out that a reset-configuration of the wireless interface fixes all the troubles. Of course the link must be good! Otherwise raise the hw-retries once again and it becomes magically more stable, but with a awful jitter.
My links are all deployed with hw-retries=4 and if it doesn't work i go to fix it and replace with a better link.

Re: Nstreme test package

Posted: Tue Apr 13, 2010 1:48 am
by rmichael
I'm deploying wireless-test since the first release and i love it,
it tolerates less the bad links (that's why you say the normal package is more reliable), but those links are working bad anyway. If you raise the hw-retries, you get the same result as the normal package.
About 4.x, i had some problems at the beginning. i found out that a reset-configuration of the wireless interface fixes all the troubles. Of course the link must be good! Otherwise raise the hw-retries once again and it becomes magically more stable, but with a awful jitter.
My links are all deployed with hw-retries=4 and if it doesn't work i go to fix it and replace with a better link.
Thanks for the info. I did not think wireless config reset would work sice I upgraded boards to v4.x before configuring any wireless on them but I'll try it. I get unusable link ~500kbps with nstreme v4.5 on. Wereas I get ~12-20Mbps, >90 CCQ with nstreme turned off (not bad for 900MHz) I don't really need nstreme at this time (noise floor @89dB) however, we got MT product because of nstreme and frame protection so it better deliver.

Btw, my links work best with hardcoded ack timeout and I suspect auto ack feature of nstreme makes it unstable.

Re: Nstreme test package

Posted: Tue Apr 13, 2010 11:30 pm
by 0ldman
4.x didn't make unstable links worse. -77 @ 24mb 90% CCQ dropped to 11mb with 20% CCQ. I wouldn't call -77 an unstable link. This is after the interface reset and reconfigure.

Also, hw retries made no difference on 3.30 vs 3.30 test for me. I set them to 15. Got a few customers I need to swap to 900MHz.

I've got a few customers on 4.x that work fine, most do not. I am running 4.3 on two different AP's. Both seem to work fine. It seems the difference is in the client side and I don't know enough about it to tell you what the difference is.

Re: Nstreme test package

Posted: Mon Apr 19, 2010 4:39 am
by rmichael
So far I tested (WDS) AP v3.30 +wireless test and Stations v4.5 and results are the same as with v4.5 for all - as long as SNR is > 28 link is stable, speed around 35Mbps. Lower SNR requires unchecking higher rates otherwise link disconnects during large data transfers. Logs show either no pooling or excessive data loss as a reason for disconnect.

PS.: interestingly, when I reset the wireless settings on AP (after moving from v4.5 to v3.3), as tgrand suggested, and CCQ went from 30 to 100%...