EIGRP "Query" & "SIA Query" Packet

[CiscoIsInYou]

Hello,

Since in

EIGRP a Query & a  SIA Query

are 2  different packet types, would really want to know the differentiation  between

EIGRP SIA Query & EIGRP Query 

and when either of these is  used by the Routers.

I went through the documentation & books only to confuse me more

Thanks.

[paul driver]

Hi, basically a eigrp SIA is a state when a router running eigrp has lost a connection to a neighbour and hasnt received a response to a query for that lost route in the allocated time = 180 secs. when this time expires the router removes the eigrp neighbour and logs a SIA event.

res

Paul

[Richard Burts · Accepted solution]

When an EIGRP router has had a route to a prefix in its forwarding table and then loses that route, then the router will send a Query message to its neighbors. The router runs a timer waiting for the query response.  When the neighbor receives the query it determines whether it can respond immediately or if it must forward the query to its neighbors. A router that has forwarded a query to its neighbors can not respond to the original query until it has received a response from all of its neighbors. If the timer expires at the originating router, perhaps because the neighbor is still wait for a response to a query that it forwarded, then the originating router will tear down the neighbor relation because it is Stuck In Active (thus the phrase SIA).

In earlier designs of EIGRP there was no way for a router to indicate that it was still waiting for response to a forwarded query and sometimes the neighbor relationship would be terminated even though both neighbors were really still communicating effectively. So Cisco added a feature in EIGRP such that after sending the original query when it lost a route then the originating router will still wait for responses and will still run the timer. But with the new feature the originating router can send a SIA Query to its neighbors before the timer expires. A neighbor may be waiting for a response to a forwarded query and can not respond to the original query but it is able to respond to the SIA query. Now the originating router knows that it must still wait for a response before it can resolve whether there is a new path to the destination, but it no longer needs to tear down the neighbor relation with the neighbor router.

HTH

Rick

[CiscoIsInYou]

Thank you pdriver for replying.

Thanks a lot Rick. This really helped.

R1====>Query/SIA Query====>R2====>Query/SIA Query====>R3====>Query/SIA Query====>R4

Could you also please explain if there's any specific time for the SIA query(e.g. 50% of Active-Timer) and whether there are nested SIA queries too as depicted above.

Thank you.

Shehjad

[Latchum Naidu]

Hi Shehjad,

The SIA state means that an EIGRP router has not received a reply to a query from one or more neighbors within the time allotted (approximately 3 minutes). When this happens, EIGRP clears the neighbors that did not send a reply and logs a

DUAL-3-SIA

error message for the route that went active.

HTH
Please click on the correct answer if this answered your question.
Regards,
Naidu.

[Richard Burts · Accepted solution]

Shehjad

Yes there is a time for the SIA Query and it is one half of the active query timer.

Yes I believe that nested SIA Query is supported.

There was an interesting presentation on EIGRP which included some discussion of the SIA Query at the recent Cisco Live. If you have access to Cisco Live presentations (perhaps through Cisco Live Virtual) you might want to look at the session Care and Feeding of EIGRP Networks.

HTH

Rick

[Jon Marshall]

Just to add to Rick's excellent posts. The main tool to use to limit SIAs in an EIGRP network is summarisation. If you have a multi-site network and you do not summarise then an EIGRP query can extend across the entire network and you are far more likely to experience SIAs.

If you summarise each site then you limit the scope of the queries which can make a significant difference in the amount of SIAs experienced.

I have seen SIAs in an EIGRP network usually because one router is heavily overloaded and does not have the resources to respond in time.

Jon

[CiscoIsInYou]

Thanks a ton Rick. You rock, 5 points to Griffindor

I saw the presentation "Care and Feeding of EIGRP Networks" & found it to be really helpful.

Last but not the least, want to thank pdriver Latchum & Jon too for their ever helpful comments.

[dslice]

I'm glad you found "The Care and Feeding of EIGRP Networks" helpful. :^)

Don. (presenter of that session)

[CiscoIsInYou]

Thank you very much Don for such insightful presentation.

R1====>Query/SIA Query====>R2====>Query/SIA Query====>R3====>Query/SIA Query====>R4

Although would like to know if such nested SIA queries are suppoorted. You also mentioned in your slides that SIA queries should be as close to the Origin as possible so would you also like to shed some light on this too?

Thanks

Shaz

[dslice]

Define "nested SIA Queries".  What really happens is that a Query is sent and after 1/2 the active time (1.5 min) if a Reply isn't received, and SIA-Query is sent to detemine if the peer is also still waiting on a Reply.  If he is also still waiting, he replies with an SIA-Reply.  I'm not sure what you mean by nested.

As for the second question, I said that the above process will normally make an SIA happen (if one happens) closer to the actual source of the problem (dirty link, router out of resources, etc.) because the other rotuers are using the SIA-Query/SIA-Replay process to verify the state of health between the routers that still have their timers running.  If the problem is realy between R3 and R4 above (dirty link, for example), the SIA-Query/SIA-Reply exchange would fail between them, causing that peer relationship to be reset, sending a Reply back to R2 and then R1.

[CiscoIsInYou]

By Nested SIA Query I mean that R1 will go Active on a route & will send a Query & in 1.5 SIA Query to R2. If R2 doesn't know the route it will Query to R3, if R3 doesn't knows the route it'll Query to R4. In 1.5 will R2 also send SIA Query to R3 & eventually will R3 also send SIA Query to R4?

What I understood is SIA Queries will be generated RECURSIVELY till the LEAF ROUTER if that route is not known.

Thank you for replying.

-Shaz

[dslice · Accepted solution]

It's not really nested or recursive.  Each router independently runs thru the diffusing update algorithm to reconcile the active route.  If a router receives a Query from his successor he looks for a feasible alternative in his topo table. If he has one, he responds with it. If not, he sends a Query to all of his peers (except the interface he received the Query on.) At the time of going active, the timer starts and if 1.5 minutes goes by without a Reply, an SIA-Query is sent to the peer(s) who didn't respond.  If they're still waiting on a Reply themselve, they respond with an SIA-Reply.  When their timer pops (1.5 minutes) that router will also send an SIA-Query, etc.

It appears to me that you have the process down quite well, but I'm not sure where the nested or recursive description comes from.

Don

[CiscoIsInYou]

Thanks a lot Don.

Now I can smell what EIGRP cooks

Appreciate everyone who contributed to answer the queries.

Shaz

[Kyaw Thet Naing]

When an EIGRP router loses a route to a destination and there is no feasible successor exists, the route is marked as "Active". Then, it will send EIGRP Query messages to its neighbors. The EIGRP router will set an Active timer after sending a Query to a neighbor and wait for a reply. If the neighbor does not respond within the Active timer, the router will send another Query (SIA Query) to ask the status of the original Query. If the neighbor cannot response the second SIA Query, the router will reset the neighbor relationship.