MikroTik

RouterOS v2.9.45 is out. We have implemented a fix for the wireless problems that arise recently in the forum about the reconnection of the clients. For all those who have this issue please upgrade to v2.9.45 and check the results. We added a new setting for hardware frame retries (hw-retries), by default it is '15'. If you experience problems in your wireless network you can try to set this number to '4'. Also we recommend for those clients that have this wireless client reconnect issue to set the 'preamble-mode=long' and 'hide-ssid=yes'

What's new in 2.9.45:

*) routing-test -
fixed memory leak;
inside confederations ignore routes with our AS number in AS path,
this fixes looping of routing information inside confederation;
*) calea server supports limited intercept;
*) mac address changing did not work for RB44G (realtek 8169 chip);
*) fixed bug - could not set new-tos in mangle rules from console;
*) realtek 8169 ethernet chip (rb44g) could stop when using vlans
*) user manager - fixed bug for PayPal payments with user data containing
specific characters;
*) implement hw-retries setting for wireless;

Thank you very much for fixing this issue! I appreciate it very much, as I'm sure everyone else does!

YYYYYEEEEEAAAAAAHHHHHH!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Can you provide a bit of a description as to what the "Hardware Retries" feature does? I have a few devices that hopefully will benefit from this fix, and I would just like to understand a bit more about what is actually happening with it.

Thanks,

Route.................eless client reconnect issue to set the 'preamble-mode=long' and 'hide-ssid=yes'

----
yes, oh.... my god !!!
I used it parameters, and after upgrade to v2....45, why my wirelessly [RB511s] encrease the signal from clients and more stable. and upgraded x86[centralized aaa] too, i don't have any problems.

Thanks Mikrotik and Team hard working for.

regards
Hasbullah.com
----

Can you provide a bit of a description as to what the "Hardware Retries" feature does? I have a few devices that hopefully will benefit from this fix, and I would just like to understand a bit more about what is actually happening with it.

I second that also... thx

Where is the routing-test package. Its not in the all-packages archive, nor
available as a separate download as an extend?

2.9.45 is a great update. I was able to work around the previous wireless problems, but the new .45 code came right up and connected with all my sites instantly (prims and atheros).

Thank you!

I they tambem I wanted to know more regarding the "Hardware Retries"...

Can you provide a bit of a description as to what the "Hardware Retries" feature does? I have a few devices that hopefully will benefit from this fix, and I would just like to understand a bit more about what is actually happening with it.

I also wanted to say thanks Mikrotik. This was a life saver!

BTW I upgraded a few of my links today, when you upgrade upon reboot the wireless cards were disabled. Just a warning to those that want to upgrade without being on site.

WOOOHOOOOO! Great stuff MT! Thanks for this great fix.

----
Hello again...
just put report image, coz i am very happy, and before i never have like this:
i put hw-retries=7 and normally all my clients not more than 3 hours have associated with central ap, now you can see uptime is more than before.



again, Thanks Mikrotik and Teams

regards
Hasbullah.com
----

What type cards are you guys using? I tested with an 8602 and it was the same old problem. Are you using SR2s?

What type cards are you guys using? I tested with an 8602 and it was the same old problem. Are you using SR2s?

We have already dropped SR2's from our network and put 8602+'s in due to the association problem, So far 4 towers upgraded and running better than ever. No association problems when changing the interface, and is it just me or is the signal strength 1000x better?

Example, a client with dynamic ACK 200 did report -85/-88 signal at 2Mb/s. NOW after upgrading to 2.9.45 the same client now reports -76/-80 ROCK SOLID!! this example is system wide across all of our clients on atheros driven MT towers now running 2.9.45

Top job MT team.

....... and is it just me or is the signal strength 1000x better?

Example, a client with dynamic ACK 200 did report -85/-88 signal at 2Mb/s. NOW after upgrading to 2.9.45 the same client now reports -76/-80 ROCK SOLID!! this example is system wide across all of our clients on atheros driven MT towers now running 2.9.45

Top job MT team.

I think your arithmetic is a bit out! An 8 to 9dB gain is around 8x, 1000x = 30dB gain.

Fantastic, after this upgade I (so far) have no association problems, I've (after succesfull reboot!) set the pre-amble to long, but due to law in our country, I cannot set hideSSID=true...

But every time I rebooted one tower, had problem, so far, so good. I'm very happy, thanks MT!

Ekkas

I think your arithmetic is a bit out! An 8 to 9dB gain is around 8x, 1000x = 30dB gain.

Merely figurative. The point is I have noticed some increase in gain.

Now after a day, I start to get more problems that before...
4 of my uplinks between towers lose connection. Sometimes it is connected but I cannot ping it, other times it just would not connect. Signal is good and they never gave problems before 2.45...

Every few minutes a link goes down and after 10 minutes or so it comes back on again. What is also strange is that for a few seconds, I can ping some router beyond the link that is down, but not the router itself, then a second or so later that is also down. Also sometimes pinging from Dude, I get 'ICMP errors' in the ping.
???
HELP!
Ekkas

Now after a day, I start to get more problems that before...
4 of my uplinks between towers lose connection. Sometimes it is connected but I cannot ping it, other times it just would not connect. Signal is good and they never gave problems before 2.45...

Every few minutes a link goes down and after 10 minutes or so it comes back on again. What is also strange is that for a few seconds, I can ping some router beyond the link that is down, but not the router itself, then a second or so later that is also down. Also sometimes pinging from Dude, I get 'ICMP errors' in the ping.
???
HELP!
Ekkas

please make the support output file at that time when you have this problem and send it to support@mikrotik.com

Upgraded to 2.9.45 on a number of sites - some with 1 ap and some with more ap's (cell sites).

on the single ap-sites everything is fine, but on the sites with more than one ap with same ssid (different freq), the registration table now shows the same mac more than once.

I suppose this is a programming error - if the client leaves one ap and connects to another on the same router, then it is logical that the old entry should be removed.

Regards
Kim C

Upgraded to 2.9.45 on a number of sites - some with 1 ap and some with more ap's (cell sites).

on the single ap-sites everything is fine, but on the sites with more than one ap with same ssid (different freq), the registration table now shows the same mac more than once.

I suppose this is a programming error - if the client leaves one ap and connects to another on the same router, then it is logical that the old entry should be removed.

Regards
Kim C

please make the support output file when you see this problem and then send it to support@mikrotik.com

I would like to say that ROS 2.9.45 is working great than I ever see about wireless AP comparing to older versions! Thx.

I would like to say that ROS 2.9.45 is working great than I ever see about wireless AP comparing to older versions! Thx.

I'd agree. This is possibly the most important release to date of 2.9.x. Our upgraded towers are running wireless AP better than they ever have. Rock Solid!

----
Hello again...
just put report image, coz i am very happy, and before i never have like this:
i put hw-retries=7 and normally all my clients not more than 3 hours have associated with central ap, now you can see uptime is more than before.

snip

again, Thanks Mikrotik and Teams

regards
Hasbullah.com
----

Another request for more info for HW retries.

I assume setting it at a lower setting simply means the AP and client will just consider it a loss in the physical layer quicker (after 7 tries rather than 15) and retransmit the whole thing, right?

----
no,
That value just my suggestion only, i like seven number. and that better for me too.
couse, in Javanese language seven is same 'pitu' same mean 'pitulungan'
pitulungan same mean making help in english.

best regards
Hasbullah.com
----

----
Hello again...
just put report image, coz i am very happy, and before i never have like this:
i put hw-retries=7 and normally all my clients not more than 3 hours have associated with central ap, now you can see uptime is more than before.

snip

again, Thanks Mikrotik and Teams

regards
Hasbullah.com
----

Another request for more info for HW retries.

I assume setting it at a lower setting simply means the AP and client will just consider it a loss in the physical layer quicker (after 7 tries rather than 15) and retransmit the whole thing, right?

hahaha!

So you just pulled that number out of thin air? lol, it works.

I upgraded, signal was a bit better, about 3-4dB gain, but CCQ was still a little wonky. Changed the HW retries, CCQ is a bit more stable now and my ping has gotten *much* more stable.

... did you change both ends or just AP?

I changed both sides. When I just changed the CPE, the AP pulled higher CCQ, still had the oddball high ping. When I changed both, the link got a lot better.

My signal was -89 on average, now its -87 or so. Never went below -86 before, now it'll go as low as -84. Overall, I'm impressed.

I changed both sides. When I just changed the CPE, the AP pulled higher CCQ, still had the oddball high ping. When I changed both, the link got a lot better.

My signal was -89 on average, now its -87 or so. Never went below -86 before, now it'll go as low as -84. Overall, I'm impressed.

Yeah I can vouch for the gain increase, what HW retries have you set?

I tried 7 as balimore did. I still haven't heard any info as to exactly what hardware retries does. I would assume a lower number means less time wasted on the physical layer and it just resends the packet. The increased stability would seem to back that up.

frame retry (the retry logic for failed transmissions) happens when packets from two or more stations collide with one another.
there are two models, unlimited and limited frame retries.

You may google for: "hardware-level frame retry" for more.

----
no,
That value just my suggestion only, i like seven number. and that better for me too.
couse, in Javanese language seven is same 'pitu' same mean 'pitulungan'
pitulungan same mean making help in english.

best regards
Hasbullah.com
----

----
Hello again...
just put report image, coz i am very happy, and before i never have like this:
i put hw-retries=7 and normally all my clients not more than 3 hours have associated with central ap, now you can see uptime is more than before.

snip

again, Thanks Mikrotik and Teams

regards
Hasbullah.com
----

Another request for more info for HW retries.

I assume setting it at a lower setting simply means the AP and client will just consider it a loss in the physical layer quicker (after 7 tries rather than 15) and retransmit the whole thing, right?

The standard setting for 802.11g is : L=7

----

nice...

regards
Hasbullah.com
----

The ....................d setting for 802.11g is : L=7

frame retry (the retry logic for failed transmissions) happens when packets from two or more stations collide with one another.
there are two models, unlimited and limited frame retries.

You may google for: "hardware-level frame retry" for more.

I've been searching, haven't found much. Seems a lot of hits for linux, not so many for wireless.

edit: hardware level without the dash gives better results, still reading, trying to find more info.