[quagga-dev 10315] Re: [PATCH] ospfv3-make-maxage-thread-periodic.patch

Greg Troxel gdt at ir.bbn.com
Mon Mar 4 12:11:09 GMT 2013


Christian Franke <chris at opensourcerouting.org> writes:

> This change certainly seems suitable to make the expiration of LSAs more
> robust.
>
> One thing I am still a little bit uncertain about: is it acceptable for
> a MaxAge LSA to sit around up to 60 seconds in the LSDB while it is
> already expired? (The specification of the aging mechanism for the OSPF
> LSDB states that it should be removed "immediately" when it is no longer
> required for flooding, I don't see any obvious harm with having
> the LSA there a bit longer, however I cannot prove that it will never be
> problematic)

I would say that the specification refers to a logical representation of
the database, and the bits are an implemntation detail.  It seems like a
reasonable strategy to have an entry in a physical database which is
marked such that a query "is this  item in the database" will be false
and iterating over all the items "in the database" will not include it.
That may require an is_item_valid().  That may be too complicated though.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 194 bytes
Desc: not available
URL: <http://lists.quagga.net/pipermail/quagga-dev/attachments/20130304/2945abe2/attachment-0001.sig>


More information about the Quagga-dev mailing list