http://blog.gmane.org/gmane.network.quagga.user hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://comments.gmane.org/gmane.network.quagga.user/9519 _______________________________________________ Quagga-users mailing list Quagga-users-UOy77sIEA+cAd7ICUelF/Q< at >public.gmane.org http://lists.quagga.net/mailman/listinfo/quagga-users John Lingate 2008-08-29T03:17:51 http://comments.gmane.org/gmane.network.quagga.user/9516 Just curious, how did they fail? I have one quagga based router which also does firewalling (hundreds of rules) and connection tracking. So far (3 years) no problems. Dennis Dennis Gerasimov 2008-08-29T00:32:02 http://comments.gmane.org/gmane.network.quagga.user/9512 _______________________________________________ Quagga-users mailing list Quagga-users-UOy77sIEA+cAd7ICUelF/Q< at >public.gmane.org http://lists.quagga.net/mailman/listinfo/quagga-users Tao Lin 2008-08-28T19:06:43 http://comments.gmane.org/gmane.network.quagga.user/9510 _______________________________________________ Quagga-users mailing list Quagga-users-UOy77sIEA+cAd7ICUelF/Q< at >public.gmane.org http://lists.quagga.net/mailman/listinfo/quagga-users Peter Mitter 2008-08-28T03:00:44 http://comments.gmane.org/gmane.network.quagga.user/9501 Can you recommend a specific motherboard make/model? Or tell me what chipsets/NIC makes/models should be the best? Also, what do think should be the absolute peak performance (throughput and PPS) I can expect using optimum hardware? Under normal conditions, 20-30Mbit. It's the DoS attacks which I need the best possible handling for. While they aren't too frequent, they can range from 50K to 500K PPS. We have 2 gigabit uplinks both with full-table BGP sessions. Stephen Marsh Name 2008-08-27T17:30:43 http://comments.gmane.org/gmane.network.quagga.user/9498 Hi List, I am planning on building a Quagga-based border router for my BGP network, and would like advice on what hardware (makes/models of motherboards, CPUs, NICs etc) will give the maximum performance; PPS is my main concern. My understanding is that a PCIe NIC with a quad-core CPU should be the best, but I'm not sure whether an onboard or external NIC would be best (or if indeed there is any difference). Your input would be much appreciated - thanks! Regards, Stephen Marsh Stephen Marsh Name 2008-08-27T16:39:52 http://comments.gmane.org/gmane.network.quagga.user/9492 Hi, From long time i have strange problems with quagga with ospf, that i can't find solution. Today again it happened so i decided to write mail to this group with hope that some one could help me. My problem is that ospf process looses adjacency, and after that ospf tryies to renew those adjacency and it does it corectly. But after that renew ospf process doesn't resend prefixes and it doesn't get new database from other routers. So basicly i don't have prefixes from other routers and other routers doesn't know about. I have quite a lot of routers connected but in this case the problem was with three of them. All of them are running FreeBSD. Router B and C are running FreeBSD 7.0 and Router A is running FBSD 6.3. I am using latest quagga 0.99.10. Router A ------ Router B ------ Router C 10.9.90.1 is on router C 212.127.90.245 is on router A This is log from quagga on router B Aug 25 11:13:32 rozanka ospfd[29913]: AdjChg: Nbr 62.233.250.34 on lan0:10.9.90.1: Full -> ExStart (SeqNumberMismatch) Aug 25 11:13:32 rozanka ospfd[29913]: nsm_change_state(62.233.250.34, Full -> ExStart): scheduling new router-LSA origination Aug 25 11:13:32 rozanka ospfd[29913]: AdjChg: Nbr 193.239.59.1 on vlan5:212.127.90.245: Full -> ExStart (SeqNumberMismatch) Aug 25 11:13:32 rozanka ospfd[29913]: nsm_change_state(193.239.59.1, Full -> ExStart): scheduling new router-LSA origination I don't know why we have SeqNumberMismatch errors. These Seq Numers are based on time AFAIK. On every hour on all routers there is scheduled time update from the same NTP server. I am not sure is this a problem or what cause those Seq Numer Mismatch. Any idea ? I think this is root of this problem. And neighbours are changed from FULL to Exstart state Aug 25 11:13:37 rozanka ospfd[29913]: Packet[DD]: Neighbor 193.239.59.1 Negotiation done (Slave). Aug 25 11:13:37 rozanka ospfd[29913]: Packet[DD]: Neighbor 62.233.250.34 Negotiation done (Slave). Negotiation is done ... Aug 25 11:13:37 rozanka ospfd[29913]: AdjChg: Nbr 62.233.250.34 on lan0:10.9.90.1: Loading-> Full (LoadingDone) Aug 25 11:13:37 rozanka ospfd[29913]: nsm_change_state(62.233.250.34, Loading -> Full): scheduling new router-LSA origination Aug 25 11:13:37 rozanka ospfd[29913]: interface vlan5:212.127.90.245: ospf_check_md5 bad sequence 1219655622 (expect 1219655624) Aug 25 11:13:37 rozanka ospfd[29913]: interface vlan5:212.127.90.245: ospf_read md5 authentication failed. Aug 25 11:13:37 rozanka ospfd[29913]: interface vlan5:212.127.90.245: ospf_check_md5 bad sequence 1219655623 (expect 1219655624) Aug 25 11:13:37 rozanka ospfd[29913]: interface vlan5:212.127.90.245: ospf_read md5 authentication failed. Aug 25 11:13:37 rozanka ospfd[29913]: AdjChg: Nbr 193.239.59.1 on vlan5:212.127.90.245: Exchange -> Full (ExchangeDone) And there is again problem with sequence number but after few tries it says that state has changed to full Aug 25 11:13:37 rozanka ospfd[29913]: nsm_change_state(193.239.59.1, Exchange -> Full): scheduling new router-LSA origination Aug 25 11:13:39 rozanka ospfd[29913]: Link State Update[Type2,id(10.9.90.14),ar(62.233.250.34)]: LS age is equal to MaxAge. Aug 25 11:13:39 rozanka ospfd[29913]: Link State Update[Type2,id(10.9.90.14),ar(62.233.250.34)]: LS age is equal to MaxAge. I have no idea why i have such log that LS is equal to MaxAge ? Could anybody tell what could cause this ? Aug 25 11:13:42 rozanka ospfd[29913]: AdjChg: Nbr 193.239.59.1 on vlan5:212.127.90.245: Full -> ExStart (BadLSReq) And another strange log. I don't know what it can mean Aug 25 11:13:42 rozanka ospfd[29913]: nsm_change_state(193.239.59.1, Full -> ExStart): scheduling new router-LSA origination Aug 25 11:13:42 rozanka ospfd[29913]: Packet[DD]: Neighbor 193.239.59.1 Negotiation done (Slave). Aug 25 11:13:42 rozanka ospfd[29913]: AdjChg: Nbr 193.239.59.1 on vlan5:212.127.90.245: Loading -> Full (LoadingDone) Aug 25 11:13:42 rozanka ospfd[29913]: nsm_change_state(193.239.59.1, Loading -> Full): scheduling new router-LSA origination Aug 25 11:13:44 rozanka ospfd[29913]: Link State Update[Type2,id(10.9.90.14),ar(62.233.250.34)]: LS age is equal to MaxAge. Aug 25 11:13:44 rozanka ospfd[29913]: Link State Update[Type2,id(10.9.90.14),ar(62.233.250.34)]: LS age is equal to MaxAge. Aug 25 11:13:44 rozanka ospfd[29913]: Link State Update[Type2,id(212.127.90.246),ar(193.239.59.1)]: LS age is equal to MaxAge. Aug 25 11:13:47 rozanka ospfd[29913]: Link State Update[Type2,id(10.9.90.14),ar(62.233.250.34)]: LS age is equal to MaxAge. Aug 25 11:13:49 rozanka ospfd[29913]: Link State Update[Type2,id(10.9.90.14),ar(62.233.250.34)]: LS age is equal to MaxAge. And again thouse strange MaxAge logs. And after that in vtysh i see that there is FULL state to router A and router C so i should have all prefixes from thouse routers in routing table. But i don't. sh ip ospf neighbor Neighbor ID Pri State Dead Time Address Interface RXmtL RqstL DBsmL 62.233.250.34 1 Full/DR 31.507s 10.9.90.14 lan0:10.9.90.1 0 0 0 193.239.59.1 253 Full/DR 32.692s 212.127.90.246 vlan5:212.127.90.245 0 sh ip route ospf show me nothing !! There is any single one prefixes from other routers. And router A and C doesn't see prefixes from router B. So basicly i have FULL adjacency with router A and C but i don't have any prefixes in routing table. sh ip ospf neighbor detail of router A: Neighbor 193.239.59.1, interface address 212.127.90.246 In the area 0.0.0.0 via interface vlan5 Neighbor priority is 253, State is Full, 82 state changes Most recent state change statistics: Progressive change 31m34s ago Regressive change 31m34s ago, due to BadLSReq DR is 212.127.90.246, BDR is 212.127.90.245 Options 2 *|-|-|-|-|-|E|* Dead timer due in 36.057s Database Summary List 0 Link State Request List 0 Link State Retransmission List 0 Thread Inactivity Timer on Thread Database Description Retransmision off Thread Link State Request Retransmission on Thread Link State Update Retransmission on sh ip ospf interface vlan5 on router B: vlan5 is up ifindex 6, MTU 1500 bytes, BW 0 Kbit <UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> Internet Address 212.127.90.245/30, Broadcast 212.127.90.247, Area 0.0.0.0 MTU mismatch detection:enabled Router ID 62.233.177.26, Network Type NBMA, Cost: 10 Transmit Delay is 1 sec, State Backup, Priority 1 Designated Router (ID) 193.239.59.1, Interface Address 212.127.90.246 Backup Designated Router (ID) 62.233.177.26, Interface Address 212.127.90.245 Multicast group memberships: OSPFAllRouters Timer intervals configured, Hello 10s, Dead 40s, Wait 40s, Retransmit 5 Hello due in 0.047s Neighbor Count is 1, Adjacent neighbor count is 1 Could someone point me what is wrong with this. Ok i understand that there is seq mismatch and ospf prcesses are trying to connect again and they do but why there is no database update after that. After 26 minutes quagga put something like this in log: Aug 25 11:39:15 rozanka ospfd[29913]: Link State Update[Type2,id(10.9.90.14),ar(62.233.250.34)]: LS age is equal to MaxAge. Aug 25 11:39:15 rozanka ospfd[29913]: Link State Update[Type2,id(10.9.90.14),ar(62.233.250.34)]: LS age is equal to MaxAge. Aug 25 11:39:15 rozanka ospfd[29913]: Link State Update[Type4,id(193.239.59.226),ar(193.239.59.230)]: LS age is equal to MaxAge. Aug 25 11:39:15 rozanka ospfd[29913]: Link State Update[Type4,id(193.239.59.226),ar(193.239.59.230)]: LS age is equal to MaxAge. Aug 25 11:42:49 rozanka ospfd[29913]: Link State Update[Type2,id(10.9.90.14),ar(62.233.250.34)]: LS age is equal to MaxAge. Aug 25 11:42:49 rozanka ospfd[29913]: Link State Update[Type2,id(10.9.90.14),ar(62.233.250.34)]: LS age is equal to MaxAge. Aug 25 11:42:49 rozanka ospfd[29913]: Link State Update[Type2,id(212.127.90.101),ar(212.127.90.101)]: LS age is equal to MaxAge. Aug 25 11:42:53 rozanka ospfd[29913]: Link State Update[Type2,id(212.127.90.101),ar(212.127.90.101)]: LS age is equal to MaxAge. Aug 25 12:48:10 rozanka ospfd[29913]: Link State Update[Type4,id(10.9.67.14),ar(193.239.59.230)]: LS age is equal to MaxAge. Aug 25 12:48:10 rozanka ospfd[29913]: Link State Update[Type4,id(10.9.67.14),ar(193.239.59.230)]: LS age is equal to MaxAge. And after that router B got all prefixes and all routes was sent to all routers. I am fighting with this problem for quite long time but i don't have any idea why this happen and how to fix this. I am not sure but there can be some bug in ospf code and i am hoping that some one could help me resolv this situation. If there is a need of some more testing and debuging feel free to ask. Thanks Bartosz Giza 2008-08-25T13:07:05 http://comments.gmane.org/gmane.network.quagga.user/9486 Hi, sorry for the somewhat alarmist subject line but I'm not sure how to tag this problem. We are using Quagga 0.98.6 and I've had to make some routing changes which involve preferencing inbound prefixes from one upstream over another. Previously it was the other way around. Before: provider A local preference = 95 provider B local preference = 80 After: provider A local preference = 95 provider B local preference = 96 We are receiving a full feed from both upstreams, which due to our filters ends up being about 260000 prefixes. When I alter the local preferences and do a "clear ip bgp B.B.B.B in" for provider B, bgpd tells zebra to pull all of the routes from provider A and insert all the routes from provider B. I'm not sure how it does this but I believe on a route-by-route basis rather than pulling all 260000 then inserting all 260000. During this operation the zebra process was using 100% CPU time, and the vtysh terminal stopped responding. I couldn't get another active vtysh terminal either. Another iBGP peer session with this particular peer went to the Idle state so I can only assume that bgpd stopped responding altogether. Either it is waiting for zebra to finish the operation, or it got stuck somewhere else. Even after zebra finished and the original vtysh terminal returned, I couldn't run any vtysh commands. Nor could I open another vtysh terminal - they just hung. Either way, it is bad for us. If we want to reclassify all routes like this for whatever reason, it will mean we have to restart bgpd (like I had to today). Anyone have any ideas about this problem? Bug, or just a Quagga limitation? Oliver Hookins 2008-08-25T03:44:41 http://comments.gmane.org/gmane.network.quagga.user/9485 _______________________________________________ Quagga-users mailing list Quagga-users-UOy77sIEA+cAd7ICUelF/Q< at >public.gmane.org http://lists.quagga.net/mailman/listinfo/quagga-users Forums - VPNVYrtual 2008-08-23T13:00:23 http://comments.gmane.org/gmane.network.quagga.user/9484 _______________________________________________ Quagga-users mailing list Quagga-users-UOy77sIEA+cAd7ICUelF/Q< at >public.gmane.org http://lists.quagga.net/mailman/listinfo/quagga-users Forums - VPNVYrtual 2008-08-23T12:56:50 http://comments.gmane.org/gmane.network.quagga.user/9483 _______________________________________________ Quagga-users mailing list Quagga-users-UOy77sIEA+cAd7ICUelF/Q< at >public.gmane.org http://lists.quagga.net/mailman/listinfo/quagga-users Forums - VPNVYrtual 2008-08-23T12:38:37 http://comments.gmane.org/gmane.network.quagga.user/9482 _______________________________________________ Quagga-users mailing list Quagga-users-UOy77sIEA+cAd7ICUelF/Q< at >public.gmane.org http://lists.quagga.net/mailman/listinfo/quagga-users Forums - VPNVYrtual 2008-08-23T12:28:04 http://comments.gmane.org/gmane.network.quagga.user/9479 Hello, Is there benefit to using net.inet.ip.fastforwarding=1 with FreeBSD and Quagga. Does enabling fastforwarding work with Quagga? Thank you, Christopher Christopher 2008-08-20T15:39:40 http://comments.gmane.org/gmane.network.quagga.user/9476 Hi, I've stumbled upon a scenario where I need to disguise one of my ASn. Scenario: My core-speaker feeds my edge-speaker with my assigned IP-range. The edge-speaker on it's side feeds to ISP. The problem that appears is that the ISP sees both my internal and external ASn and ends up invalidating the feed as it's original origin is from an unknown ASn. Now, is there some way to disguise my internal ASn with the external? Thanks /Jonathan Petersson Jonathan Petersson 2008-08-19T21:50:14 http://comments.gmane.org/gmane.network.quagga.user/9474 _______________________________________________ Quagga-users mailing list Quagga-users-UOy77sIEA+cAd7ICUelF/Q< at >public.gmane.org http://lists.quagga.net/mailman/listinfo/quagga-users Ravi Singla 2008-08-19T05:48:29 http://comments.gmane.org/gmane.network.quagga.user/9470 Hello, I am new to Quagga and I am currently running 0.98.6 on FreeBSD 7.0. I have zebra, ospf, and bgp running...but bgp is not working. Static routing and OSPF has been working great for over a week. But for BGP, I just tried establishing a BGP session to an upstream provider with a Juniper and the neighbor session just cycles between Connect -> Active -> Connect. In show ip bgp neighbor I do not see any counters incrementing. This is what I see in BGP debugs.... 2008/08/18 01:55:49 BGP: 216.85.157.1 [FSM] Non blocking connect waiting result 2008/08/18 01:56:44 BGP: Performing BGP general scanning 2008/08/18 01:57:04 BGP: 216.85.157.1 [Event] Connect failed (Operation now in progress) 2008/08/18 01:57:04 BGP: 216.85.157.1 [FSM] TCP_connection_open_failed (Connect->Active) 2008/08/18 01:57:04 BGP: 216.85.157.1 went from Connect to Active 2008/08/18 01:57:44 BGP: Performing BGP general scanning 2008/08/18 01:58:44 BGP: Performing BGP general scanning 2008/08/18 01:59:04 BGP: 216.85.157.1 [FSM] Timer (connect timer expire) 2008/08/18 01:59:04 BGP: 216.85.157.1 [FSM] ConnectRetry_timer_expired (Active->Connect) 2008/08/18 01:59:04 BGP: 216.85.157.1 went from Active to Connect 2008/08/18 01:59:04 BGP: 216.85.157.1 [Event] Connect start to 216.85.157.1 fd 12 2008/08/18 01:59:04 BGP: 216.85.157.1 [FSM] Non blocking connect waiting result 2008/08/18 01:59:44 BGP: Performing BGP general scanning 2008/08/18 02:00:19 BGP: 216.85.157.1 [Event] Connect failed (Operation now in progress) 2008/08/18 02:00:19 BGP: 216.85.157.1 [FSM] TCP_connection_open_failed (Connect->Active) 2008/08/18 02:00:19 BGP: 216.85.157.1 went from Connect to Active This is my bgp config.... router bgp 29724 bgp router-id 216.85.157.2 bgp log-neighbor-changes network 63.247.192.0/19 neighbor timewarner peer-group neighbor timewarner remote-as 4323 neighbor timewarner description BGP to TimeWarner neighbor timewarner soft-reconfiguration inbound neighbor 216.85.157.1 peer-group timewarner I've double checked and MD5 is not used on this connection (and not compiled into this version of quagga). Any ideas on what I am missing here? Thank you, Christopher Christopher 2008-08-18T08:04:37 http://comments.gmane.org/gmane.network.quagga.user/9468 Hi, Basicly i want to distribute default route into ospf when my bgp connection is up with my ISP. When there is link failure and my bgp session terminates i would like to stop distributing default route into ospf. I have tried some setups with route-map and checking for specific prefix in routing table but it didn't worked. Could anyone point me how to do this? Thx Bartosz Giza 2008-08-15T21:03:53 http://comments.gmane.org/gmane.network.quagga.user/9463 Hello, I run BGP to connect with one AS to two transit providers. Each transit providers is connected to a quagga router, total two routers. OSPF is used between the two quagga routers. I can announce a net with the two providers, no problem. I have added "Return Path filters" found at http://wiki.quagga.net/index.php/Main/TipsNTricks net.ipv4.conf.all.rp_filter = 0 net.ipv4.conf.lo.rp_filter = 0 net.ipv4.conf.default.rp_filter = 0 But when an incoming packet is coming from ISP-1 and the VLAN is on quagga-2 the outgoing packets is going out through ISP-2. This works most of the time. But for protocols like FTP it doesnt work. Does anyone have an idea how to get the packets in and out of the same quagga router even if the VLAN is on the other quagga router? /JB JB 2008-08-15T05:41:28 http://comments.gmane.org/gmane.network.quagga.user/9462 Quagga does not deal with queue congestion at all, because it doesn't handle the forwarding of network traffic. The Quagga daemons use routing protocols to talk with other routers, and use the routes they learn to populate the routing tables of the host operating system with routes. The host operating system (usually GNU, Linux or one of the BSDs) is the component which actually handles IP traffic, routes packets between interfaces, queues packets, and, therefore, manages packet queues and congestion thereof. Assuming you're running Quagga on a Linux based system, take a look at the network sections of the Linux kernel source code: http://www.kernel.org/ Charles Briscoe-Smith 2008-08-14T23:53:13 http://comments.gmane.org/gmane.network.quagga.user/9461 _______________________________________________ Quagga-users mailing list Quagga-users-UOy77sIEA+cAd7ICUelF/Q< at >public.gmane.org http://lists.quagga.net/mailman/listinfo/quagga-users Jose Rodriguez 2008-08-14T22:39:25 http://comments.gmane.org/gmane.network.quagga.user/9460 _______________________________________________ Quagga-users mailing list Quagga-users-UOy77sIEA+cAd7ICUelF/Q< at >public.gmane.org http://lists.quagga.net/mailman/listinfo/quagga-users Swati Ghantiwala 2008-08-14T18:09:50 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.network.quagga.user