watchdog kill dead nstreme

2.8.26 (I know nobodys interested in it anymore)

may/04/2005 14:32:24 wlan: watchdog kill dead nstreme, 00:xx:yy:zz:17:D8
may/04/2005 14:32:24 wlan: 00:xx:yy:zz:17:D8 disconnected
may/04/2005 14:32:27 wlan: 00:xx:yy:zz:17:D8 connected

just to let you know. Intel 440BX platform, Celerons 400MHz and around, Intel 815 with P3/866, enough memory, almost no uptime :slight_smile:

uptime: 5d9h57m46s
free-memory: 468708 kB
total-memory: 515480 kB
cpu: Pentium
cpu-frequency: 863 MHz
cpu-load: 10
free-hdd-space: 28267 kB
total-hdd-space: 60540 kB
write-sect-since-reboot: 3470
write-sect-total: 7432

3km ptp line, otherwise working fine. Three second interruption, then the line lives again. This happens ca. 3 times a day, it is NOT depending on load, no power problems, no temperature problems, routers do not reboot nor experience instability. Everything is perfect.

I couldn’t find dependency or pattern. 2.8.22 was not doing this, I didn’t try versions between.

I see the same issue on a 7 mile link with 2 integrated boards and CM9. The link works fine, but there are these log the show the connection/disconnections.


may/08/2005 20:42:32 wlan2: watchdog kill total timeout, ***:7C:90
may/08/2005 20:42:32 wlan2: ***:7C:90 disconnected
may/08/2005 20:42:35 wlan2: ***:7C:90 connected

exactly the same - three second, three timeout on pings and the link is back again.

so far, I found nothing on the web and nobody replied to this post as you can see. What I did today was disabling periodic-calibration. Other than that, I couldn’t imagine anything else on my system what could be related to this - then the only option available would be problem in software.

How far apart are your nodes? I am thinking this might be a ACK-timeing issue.

Mine was set a dynamic, when I click on status, I see it is 25. The handy chart in the Docs suggest for that my 5km shot, I should be using 52. So I am now playing with that.. Ill post how it goes.

I have short links, up to 3.3km (=2 miles). ACK timing is not the issue.

bump…


may/29/2005 05:33:10 router: watchdog kill dead nstreme, 00:xx:xx:xx:17:D8
may/29/2005 05:33:10 router: 00:xx:xx:xx:17:D8 disconnected
may/29/2005 05:33:11 router: unauth or missing data sender,
00:xx:xx:xx:17:D8
may/29/2005 05:33:11 router: deauth data sender, 00:xx:xx:xx:17:D8
may/29/2005 05:33:11 router: unauth or missing data sender,
00:xx:xx:xx:17:D8
may/29/2005 05:33:11 router: deauth data sender, 00:xx:xx:xx:17:D8
may/29/2005 05:33:11 router: unauth or missing data sender,
00:xx:xx:xx:17:D8
may/29/2005 05:33:11 router: deauth data sender, 00:xx:xx:xx:17:D8
may/29/2005 05:33:15 router: 00:xx:xx:xx:17:D8 connected
may/29/2005 05:34:00 router: watchdog kill dead nstreme, 00:xx:xx:xx:17:D8
may/29/2005 05:39:28 router: watchdog kill dead nstreme, 00:xx:xx:xx:17:D8
may/29/2005 05:39:28 router: 00:xx:xx:xx:17:D8 disconnected
may/29/2005 05:39:31 router: 00:xx:xx:xx:17:D8 connected
may/29/2005 05:45:16 router: watchdog kill dead nstreme, 00:xx:xx:xx:17:D8
may/29/2005 05:45:16 router: 00:xx:xx:xx:17:D8 disconnected
may/29/2005 05:45:19 router: 00:xx:xx:xx:17:D8 connected
may/29/2005 10:36:30 traffic logger configuration changed by admin
may/29/2005 10:37:59 router: watchdog kill dead nstreme, 00:xx:xx:xx:17:D8
may/29/2005 10:37:59 router: 00:xx:xx:xx:17:D8 disconnected
may/29/2005 10:38:01 router: 00:xx:xx:xx:17:D8 connected


interface-type=Atheros AR5212
chip-info=“mac:0x5/0x6, phy:0x41, a5:0x36, a2:0x0”
tx-power-control=yes ack-timeout-control=yes alignment-mode=yes
virtual-aps=yes noise-floor-control=yes scan-support=yes burst-support=yes
nstreme-support=yes default-periodic-calibration=disabled


/interface wireless:
0 R name=“router” mtu=1500 mac-address=00:xx:xx:xx:5D:2A arp=enabled
disable-running-check=no interface-type=Atheros AR5212
radio-name=“router” mode=ap-bridge ssid=“router”
frequency=2437 band=2.4ghz-g-turbo
scan-list=default-ism
rate-set=configured supported-rates-b=1Mbps,2Mbps,5.5Mbps,11Mbps
supported-rates-a/g=6Mbps,9Mbps,12Mbps,18Mbps,24Mbps,36Mbps
basic-rates-b=1Mbps basic-rates-a/g=6Mbps max-station-count=3
ack-timeout=dynamic tx-power=default noise-floor-threshold=default
periodic-calibration=disabled burst-time=disabled fast-frames=yes
dfs-mode=none antenna-mode=ant-a wds-mode=disabled
wds-default-bridge=none wds-ignore-ssid=no
update-stats-interval=disabled default-authentication=yes
default-forwarding=no hide-ssid=no 802.1x-mode=none
disconnect-timeout=3s on-fail-retry-time=100ms


nstreme:
0 name=“router” enable-nstreme=yes enable-polling=yes
framer-policy=exact-size framer-limit=4000


what does this mean and how can it be solved? We tried all combinations of v2.8.26 and v2.8.22 software on AP and client and I’m getting pretty angry about this. 3km link, point to point, -60dB signal but I have many other links - shorter and longer - that do EXACTLY THE SAME. The traffic stops, radio de-associates, then associates again and everything starts working again, but this is very annoying and TOO FREQUENT. Different ATX machines, different boards, different cards, totally stable systems (they can have 80 days uptime, but this issue is not uptime dependant, currently this log is posted from system rebooted at midnight, so it took 5 hours to appear).

In order to see this in logs, you have to turn ON wireless info in /system logging facility (set wireless-info echo=yes). As you can see, I turned off periodic calibration, but it didn’t help. Other than that, I have switched off default forwarding, set maximum station count to 3 (default: 2007) and switched on fast-frames. Whats ahead of me is turning OFF polling and trying to mess up with framer policy (it was broken some versions ago, when you set dynamic, it reported exact or vice versa).


Anyone met this before? I guess no solution is available, because this thread is totally dead.

thnx, mp3turbo.

I am still here with you, although I have made 2 of my links work by changing the frequency on one, and increasing the signal by using a bigger dish.

thnx surfnet !

Changing frequencies didn’t help me, that was the first thing I’ve done. I could change RF part (antenna, pigtail, cable), but it doesn’t make sense for me as signal level is not changing and overall quality readings (ccq etc) are 75.

I’m also having a problem with this. I have about 3 wireless clients out of 8 that are doing this. Anyone ever figure out what’s going on?

Zach

I wonder if this problem is related?

http://forum.mikrotik.com/t/wireless-disconnects-exactly-every-3-hours/2532/1

Hitek

Did you ever find out the answer to the problem ?
FW upgrade ?