slow redistribution between OSPF processes

[martap]

Hi All,

Looking for some advice, will appreciate if you can help.

I run two OSPF processes and redistribute prefixes between them. I also run BFD to allow for faster failure detection which works great. What I notice is that in case of a link down, bfd immediately signals OSPF who then withdraws the prefixes. All good so far. However, the other ospf process only seem to "notice" the change about 15 sec later... Redistribution seems slow. Is there anyway to speed this up?

[paul driver]

Hello

apologies not replying sooner I don’t have any notification when a post is replied to so it’s hard to keep track?

so to summarise when you drop any related redistributed prefix ospf doesn’t remove it from for upto 15 seconds ? Is this from the rib or the database or both?

[martap]

Absolutely no problem, Paul

From both… DB & RIB.

[paul driver]

Hello

I think what you are seeing the synchronization of the database

When a external route is removed, ospf is aging out the route and it processing process 1 first, I did a simple test and got the same result as yourself below you can see two entries in the D/B just before i removed route 4.4.4.4 then after then you can see ospf maxage kicking in when this route is removed and it processing each opsf process individually.

sh ip ospf database external 4.4.4.4 | in Ad|LS
Routing Bit Set on this LSA
LS age: 19
LS Type: AS External Link
Advertising Router: 3.3.3.3 < my router connected to ospf 2
LS Seq Number: 80000001
Forward Address: 10.1.34.4

LS age: 18
LS Type: AS External Link
Advertising Router: 1.1.1.1 < my router connected to ospf 1
LS Seq Number: 80000001
Forward Address: 0.0.0.0

*Mar 1 01:10:57.075: Stop timer for 4.4.4.4 1.1.1.1 5 3

sh ip ospf database external 4.4.4.4 | in Ad|LS
Delete flag is set for this LSA
LS age: MAXAGE(3604)
LS Type: AS External Link
Advertising Router: 3.3.3.3
LS Seq Number: 80000002
Forward Address: 0.0.0.0

#sh ip ospf database external 4.4.4.4 | in Ad|LS
Delete flag is set for this LSA
LS age: MAXAGE(3606)
LS Type: AS External Link
Advertising Router: 3.3.3.3
LS Seq Number: 80000002
Forward Address: 0.0.0.0

#sh ip ospf database external 4.4.4.4 | in Ad|LS
Delete flag is set for this LSA
LS age: MAXAGE(3607)
LS Type: AS External Link
Advertising Router: 3.3.3.3
LS Seq Number: 80000002
Forward Address: 0.0.0.0

#sh ip ospf database external 4.4.4.4 | in Ad|LS
Delete flag is set for this LSA
LS age: MAXAGE(3608)
LS Type: AS External Link
Advertising Router: 3.3.3.3
LS Seq Number: 80000002
Forward Address: 0.0.0.0

sh ip ospf database external 4.4.4.4 | in Ad|LS
Delete flag is set for this LSA
LS age: MAXAGE(3609)
LS Type: AS External Link
Advertising Router: 3.3.3.3
LS Seq Number: 80000002
Forward Address: 0.0.0.0

*Mar  1 01:11:04.567: Stop timer for 4.4.4.4 3.3.3.3 5 3
sh ip ospf  database external 4.4.4.4 | in Ad|LS

[martap]

I see… but the result is that for 15-20 sec, the traffic is blackholed which is what I’m trying to avoid. I assumed BFD would help me get quicker convergence (<1sec).

[paul driver]

Hello

BFD is a link detection feature that detects failure faster than say ospf would, So i not sure it will have the desired effect your trying accomplish.

Ospf has a similar feature using the dead-interval 
int x/x
ip ospf dead-interval minimal hello-multiplier xx

However If you redistributing between the two ospf process anyway wouldn't be viable to just have the one ospf process and two areas I would envisage it would be somewhat faster ospf convergence?

[martap]

Is there anyway to influence the MAXAGE timer? 

The reason I'm using 2 processes, is to get more control over the prefixes that I need redistributed.. with just 2 areas I'm severely limited with OSPF unfortunately.

thanks for your help so far, Paul. Much appreciated.

[paul driver]

Hello

---

@martap wrote:

Is there anyway to influence the MAXAGE timer? 

---

As far as I am aware no there isn't as ospf is an open vendors protocol I think this need to be consistent across the board