Hi,
We have an issue with OSPF.


OSPF :
interfaces network type PTP
CCR1009 ---- AF5x#1 ---- AF5x#2 ---- CRS125 ---- RB960
10.10.10.0/29


Netwatch pinging :
AF5x#1
AF5x#2
RB960

RB and CRS == V6.4.1.4
CCR == V6.4.2
AF5x == V3.2.3


AF5x = never had issues since at least 1 year. Only since OSPF is running. UPTIME 224days
Capacity : 300MBPS and only use 100mbps MAX peak
CCR1009 and RB960 don't record any tx errors or link down



After 1 day or 2, the path goes down for a about a minute then come back to full. When the link Is down Newatch notify me and we received an eamil that AF5x, crs125 and rb960 are down.
While the down time, I can acces The AF5x#1 and see that the wireless link is UP and running but cant acces the remote af5x. Thats sound strange for me.
When I do the test, I am connected to my CCR1009's VPN.


Any idea? Any suggestion? Need more information?


Thank you very very much

Hi ubikrotik:

You'll need to double check if you have configured the entire parameters with OSPF, in order to keep your neighbors up, particularly with the timers (hello / dead interval ), i've worked with AF4x, AF5x, AF24x and i've never seen that behavior, you'll need to upgrade RouterOS and AirOS and check if everything is fine, then we'll need to see some logs of that OSPF issue

yes, routeros and routerboard have the latest fw.

in every routers i can see all of my lsa and neighbors. hello, dead interval are set to default. i got a new log last days :


OSPF - init bit set in middle of an exchange

PTP on ospf type? You dont have that. try to switch to broadcast or nbma.

broadcast do the same. thats why we switch to ptp.

Envoyé de mon Pixel XL en utilisant Tapatalk

we also tried authentification to simple. no cha ges.

Envoyé de mon Pixel XL en utilisant Tapatalk

Quick question, are you able to reach via Mac-telnet your remote router?? Or do you see that remote router as an LLDP Neighbor (/ip neighbors print) ??

It's hard to say. When it happens I don't have enough time to remote acces my main router and do that. The only thing I can say, is that I can't ping the remote router from the main router. (of course if ospf is down). My wireless link Is very stable and strong. No worries about the link.


any idea :S?

Interesting, you may be able to reach your remote wan interface from the local router (WAN1 -> WAN 2), do you have your timers configured with default values?

exactly, and also unable to reach the remote PTP radio. But it's still up and running because, when i'm logged in the first PTP radio, I can see the other one running and connected. The link uptime does not reset too. So im sure that the PTP link is good without any problems.

We tried NBMA without succes.

timers have defaults values yes.

realy crazy problem.

If the link is good but I can't acces the remote radio, switch and router there is something that block me ... but what? this thing also block the route to this router.

thank you for your help.

regards.

Got you, that seems like the AirFiber is blocking something in the middle, are they configured as bridge? Do you have any diagram that you can upload here? I just need to understand the Layer3 / Layer2 diagram between both routers and both AirFibers... do you have any log related to bad ethernet negotiation?

Check that WDS is enabled on BOTH sides of the ubnt link.

You dont have anny more setup than you provided in this post? - Eg a way for ospf to reach first router without going trough your airfiber. I have seen this before - if a router behind for some reasion provide the /xx routes tor first router but make it self the gateway.

If not. there is also a possiblility that you are missing some of your packets troug airfiber. I do not know arifiber to well, i stoped using ubnt equipment years ago due to a lot of problems.mainly l2mtu and the drop that fill the glass - a firmware upgrade blocking ethernet access to devices. But that said - i have had a lot of licensed equipment etc, and running ospf between MT routers in each ends - and there is always. Always a drop making ospf to goes down. it might be wireless issue - or cable. signal could be fine - but still dropped packets.one way is to do a full wlan research and calculation to your link. If a calculator says signal should be -55 and you have -60 - there is a problem. It might be noice, aligment - or something else. But still +/- 3dbm - you have a problem that might couse problem. Also make sure cable is ok - and that you are not maxing out interface speed/traffic. Then packed could be dropped. If you can reach radio1 and not radio2 - i can tell you 99,9% for sure - there is something wired going on between the radioes. But the 0,1% left, is how ubnt make psudo interfaces etc to have traffic trough wlan and ethernet. ITs a big "hack", and in some rare cases make trobble, also if yo do something manualy in config to ubnt radioes - it will make trobles. Eg manualy add vlan into its some kind of psudo interfaces - connecting to a br interface etc. (what makes my interface dissable after firmware upgrade) Im like 99,9% sure - if you put a good ethernet cable between routers with ospf - you will not have the problem. Then - search the radioes (airfiber) for problem. and dont say like - ah. signal seems nice- its not signal. KNOW that signal is what calulator say etc etc. A lot of problem is solved, just by error handling stuff that ppl think, and say----- is not the problem.

well, I can share to you the diagram. But this is really simple.

our main router has a cable that is connected to an airfiber. At the other end, we have the remote airfiber, a cable, a switch a cable and the router #2.


/29 for all of this devices
eg:
10.10.10.1 is the main router
10.10.10.2 is the first af5x
10.10.10.3 is the second af5x
10.10.10.4 is the switch
10.10.10.5 is the router#2.

No other routes possible to reach router #1


yes signal is good only 1 pckts lost for a 24hrs continous pingging. min:0 avg:2 max:28

I dont know why Ive lost 1 pckts. but i think its acceptable.

no error on cable no cable negotiation log, no disconnection. no uplink downtime.
steady and stable TP.



today we removed the switch between and upload the last fw to each af5x
no more drops since 24hrs....


ill keep you in touch.





Envoyé de mon Pixel XL en utilisant Tapatalk

dup, sorry

Change your P2P frequency.

10.10.10.4 is the switch - what kind of switch=? Doeas it use rstp? . Then dissable it on all devices. You dont need it if you have control to your networkm and ypu dont have loop.

10.10.10.4 is the switch - what kind of switch=? Doeas it use rstp? . Then dissable it on all devices. You dont need it if you have control to your networkm and ypu dont have loop.

crs112


rstp? not sure?

Envoyé de mon Pixel XL en utilisant Tapatalk

yes this is default /bridge/stp

protocol mode : rstp

on all bridge. should i use none?

Envoyé de mon Pixel XL en utilisant Tapatalk

and should i use auto isolate (port setting)?

Envoyé de mon Pixel XL en utilisant Tapatalk

Hi, just disable SPT (set to protocol=none)... but I guess that your wireless link configuration may be blocking something, according to your network diagram (everything in the same bridge), if the ospf connection goes down it shouldn't affect the reachability of your network devices between the same Layer2 domain

Hey guys i think its now fixed. Since this monday, i removed the switch and installed the lastest version in my af5x.


No more drops/down.


if it come back again I will let you know[emoji2]


thank you very much for your help.!!!


there is an other issue in our network now. If you want to take a look[emoji6]

viewtopic.php?t=134304

Cool!!! That's great!!