OSPF: Socket error: No buffer space available
Hello, Can someone pls. provide some insight into this? Am a first-time user of BIRD and OSPF. I have two machines (bird0 and bird1) running bird with a OSFP router for the same area configured and both using a non default routing table (vr1) pointing to kernel routing table 1. One of the birds config files (on bird0) also has a large number of static routes configured, also for vr1. There are 512k static routes with contain more than 100k unique prefixes. The idea is that the other birds (bird0) OSFP will get a dump of these routes. - starting up bird0 and after a while: bird0: birdc show route table vr1 | wc -l --> 116977 bird0: ip route show table 1 | wc -l --> 116973 - then starting up bird1 and after a while: bird1: birdc show route table vr1 | wc -l --> 116977 bird1: ip route show table 1 | wc -l --> 116974 I see the "bird: OSPF: Socket error: No buffer space available" error on bird1 console, BUT all routes are propagated so there does not seem to be an issue handling this many routes. So is this some type of transient condition? Or is there a deeper hidden issue here handling this volume of routes, for which i may have to tweak things? Any help here is greatly appreciated. Thanks Sid Narasimhan
On Wed, Jun 08, 2011 at 03:47:53PM -0400, Sid Narasimhan wrote:
Hello,
One of the birds config files (on bird0) also has a large number of static routes configured, also for vr1. There are 512k static routes with contain more than 100k unique prefixes. The idea is that the other birds (bird0) OSFP will get a dump of these routes.
..
I see the "bird: OSPF: Socket error: No buffer space available" error on bird1 console, BUT all routes are propagated so there does not seem to be an issue handling this many routes.
So is this some type of transient condition? Or is there a deeper hidden issue here handling this volume of routes, for which i may have to tweak things?
I am not sure, but i would guess that it is a transient problem - too many packets are generated in a moment, some of these are lost, which is fixed later by OSPF retransmission mechanism. BTW, i wouldn't recommend exporting 100k prefixes to OSPF (for example, i would guess the synchronization would be slow, how much time it took in your case?). Generally, such number of static routes could be propagated much more efficiently using BGP. But it depends on your use case. -- Elen sila lumenn' omentielvo Ondrej 'SanTiago' Zajicek (email: santiago@crfreenet.org) OpenPGP encrypted e-mails preferred (KeyID 0x11DEADC3, wwwkeys.pgp.net) "To err is human -- to blame it on a computer is even more so."
participants (2)
-
Ondrej Zajicek -
Sid Narasimhan