On Mon, Nov 16, 2015 at 6:35 AM, Paul Jakma <p...@jakma.org> wrote: > On Wed, 21 Oct 2015, Daniel Walton wrote: > > Signed-off-by: Daniel Walton <dwal...@cumulusnetworks.com> >> Reviewed-by: Donald Sharp <sha...@cumulusnetworks.com> >> --- >> bgpd/bgpd.h | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/bgpd/bgpd.h b/bgpd/bgpd.h >> index b57a62a..f9bee3c 100644 >> --- a/bgpd/bgpd.h >> +++ b/bgpd/bgpd.h >> @@ -752,7 +752,7 @@ struct bgp_nlri >> #define BGP_DEFAULT_EBGP_ROUTEADV 0 >> #define BGP_DEFAULT_IBGP_ROUTEADV 0 >> #define BGP_CLEAR_CONNECT_RETRY 20 >> -#define BGP_DEFAULT_CONNECT_RETRY 120 >> +#define BGP_DEFAULT_CONNECT_RETRY 10 >> > > 10s? > > Why not lower again? 5s would be unlikely to cuase problems for most? 2s > probably wouldn't hurt anyone either. >
The thought was that for iBGP peers it would be worth waiting long enough for the IGP to converge before we make another attempts at establishing the connection. 10s should be plenty of time for that...I don't have strong feelings against making it lower though since we could probably set it to 1s and I doubt anyone would ever notice in terms of CPU. Daniel
_______________________________________________ Quagga-dev mailing list Quagga-dev@lists.quagga.net https://lists.quagga.net/mailman/listinfo/quagga-dev