Upcoming Maintenance Alert:

The UBNT Community will be upgraded at 5pm MDT on April 25th. During this time the community forums will be set to read-only status.

Learn more

×

Add a clear ip ospf process to commits that require it.

Submitted by -
Status: New Idea

Certain changes make ospf stop working until a "clear ip ospf process" command is issued to cause a reboot of ospf and re-convergence with other routers.

 

For example a change to the ospf router-id:

 

set protocols ospf parameters router-id 10.50.14.14

 

Right now, edgeOS issues a warning after the commit (on the above router-id change) that the clear ip ospf process needs to be issued.

 

But unfortunately, by the time the message is displayed, ospf has already stopped working and the message would only be seen if you are locally connected to the router.  

 

If the ospf router-id is changed while connected to the router via OSPF, then you will loose connection to the router and never be able to issue the "clear ip ospf process" to get access back.

 

You could stack the clear ip command to the commit, but not everyone would know or remember to do that.

 

It seems that EdgeOS is already able to detect that the clear ip ospf process is required.  So instead of issuing a warning that probably won't be seen, it should just clear the process and give a message that it did so.

 

If there is some reason I'm not thinking of where you wouldn't want this to be automatic, then give a message like:

 

ospf process will be cleared in 30 seconds.  (Abort?)

 

This would allow those folks that for some reason wouldn't want it cleared yet to abort the auto clear.

 

Anyhow, I think this would save a lot of people from loosing connection to their routers and truck rolls.  Most routers running ospf will be accessed via OSPF and never see the current message. 

Comments
by
on ‎09-26-2017 07:27 AM

As discussed here ... The bigger issue is that said command clears everything, and will pull your network down, if done on a core router.

 

The problem with automating "clear ip ospf process" is that this needs to be done every time you do a OSPF change, not just Router ID, and when you run it, it will remove every route, re-initilize everything, going through hello packets, neighbors etc. and then add them back.

 

This process takes atleast 10'ish seconds, and on a production network, that results in a 'full stop'.

 

This is very ineffective, since for instance, when you do a change on a Tik, say you change/add/delete a OSPF branch, it will only affect the 'leg' your working on, and everything else is still fully functional.

 

Meaning you can re-route traffic while working on said 'leg'/'branch', and work on a live network, and add it back when ready, without effecting your entire infrastructure.

 

I'm all for a auto refresh, as long as it's not a complete 'Clear All'.

 

/Paetur