Community discussions

MUM Europe 2020
 
-headstrong-
Member
Member
Topic Author
Posts: 377
Joined: Thu Jun 05, 2008 8:04 pm
Location: South Africa

OSPF Changing states: Full to Down..why?

Mon Jan 19, 2009 3:34 pm

Hi guys

We have quite a large network and we are using OSPF routing. Every now and then some of our backbone links change state from Full to Down which causes some online applications to drop.

Some of our backbone links are grids connecting to Sectors which have clients on them.Could this be the problem? Should we use Grid <====> Grid links for OSPF???

thanks
 
User avatar
JJCinAZ
Member
Member
Posts: 473
Joined: Fri Oct 22, 2004 8:03 am
Location: Tucson, AZ
Contact:

Re: OSPF Changing states: Full to Down..why?

Wed Jan 28, 2009 10:51 pm

Yes, if you have too much packet loss or delay, the OSPF sessions can drop. Think about it, how does OSPF normally detect a dead link?

Should you use dedicated point-to-point links for backhaul? If it makes economic sense, then by all means do that.
 
-headstrong-
Member
Member
Topic Author
Posts: 377
Joined: Thu Jun 05, 2008 8:04 pm
Location: South Africa

Re: OSPF Changing states: Full to Down..why?

Fri Jan 30, 2009 12:59 pm

Thanks for your reply...There wasn't really any packet loss until the OSPF dropped. I am now statically routing some of the highsites and now there is zero packet loss....

So you saying that having a point-to-point backhaul link compared to a point-to-multi-point link will have no difference in OSPF performance?
 
Mplsguy
MikroTik Support
MikroTik Support
Posts: 226
Joined: Fri Jun 06, 2008 5:06 pm

Re: OSPF Changing states: Full to Down..why?

Fri Jan 30, 2009 5:10 pm

One potential issue that may affect OSPF in noisy point-to-multipoint environments is the way multicasts are transmitted (OSPF hellos in this case) - when AP transmits multicast (or broadcast), it is transmitted as radio multicast without retries - if that single transmission attempt is lost, frame is lost. When Station transmits multicast it is transmitted as radio unicast (to AP), therefore its delivery to AP is guaranteed. You should be able to figure out if this is the case by enabling OSPF packet logs on both sides and checking if OSPF hellos are lost.

If this is the case there are several things you can do:
- configure OSPF neighbors manually (so that unicast hellos are sent)
or
- try to avoid/reduce noise/interference on your backhaul link (e.g. by moving it to different channel and having only PtP link for your backhaul)
or
- by setting up WDS link (either AP-to-AP or AP-to-stationwds) between devices forming backhaul - traffic over WDS links is always sent as radio unicast and therefore is subject to retry algorithm.
 
-headstrong-
Member
Member
Topic Author
Posts: 377
Joined: Thu Jun 05, 2008 8:04 pm
Location: South Africa

Re: OSPF Changing states: Full to Down..why?

Fri Jul 31, 2009 12:11 am

Hi mpls guy

apologise for the delayed response...I've changed my links to WDS but one of them is an nstreme link which causes issues with WDS..

Is the below config correct with Nstreme enabled at both ends:

AP:
wlan1 in ap-bridge mode with static WDS enabled
Created a WDS interface with wlan1 as parent
Gave WDS interface an IP 192.168.0.2/30

Client:
wlan2 in station-wds with static WDS enabled
Created a WDS interface with wlan2 as parent

I cannot add an IP to the Client WDS interface, why is that? If I keep it on the wlan2 interface then the link works but is it a WDS link?
 
-headstrong-
Member
Member
Topic Author
Posts: 377
Joined: Thu Jun 05, 2008 8:04 pm
Location: South Africa

Re: OSPF Changing states: Full to Down..why?

Sun Aug 02, 2009 10:27 pm

Would creating OSPF "sham" links between the sites be useful? Yes I know that this is not the purpose of sham links but would it prevent the OSPF from changing between the states?

Who is online

Users browsing this forum: MSN [Bot] and 19 guests