From esj at cs.fiu.edu Sat Oct 1 10:43:40 2011 From: esj at cs.fiu.edu (Eric S. Johnson) Date: Sat, 01 Oct 2011 13:43:40 -0400 Subject: [Xorp-users] OSPF slow-to-converge bug fixed. In-Reply-To: Your message of "Fri, 30 Sep 2011 14:16:05 PDT." <4E863195.2040304@candelatech.com> Message-ID: <20111001174340.1224B3680002@cheetah.cs.fiu.edu> >Anyway, it appears to be fixed now. So, if anyone had any issues with >OSPF being slow to converge, please try out the top-of-tree and see >if it works better for you. I was one of those that had issues. I had tried playing with the environment variable setting you sent me a few weeks ago, but even extreme settings didn't seem to help, just to push the problem later. I didn't have time to do much more debugging myself. I am very happy to report that the git tree I just pulled and built works perfect. And ospf starts doing its thing much faster now too. (Centos 5, ospf and PIM-SM primary used features) This bug is squashed! Thanks much! E From greearb at candelatech.com Mon Oct 3 17:33:31 2011 From: greearb at candelatech.com (Ben Greear) Date: Mon, 03 Oct 2011 17:33:31 -0700 Subject: [Xorp-users] Xorp doesn't start In-Reply-To: <4DEE6EA7.5080105@ll.mit.edu> References: <80C343A5D090C24F9229C34CCD17287405F37ABF@usnjpar1blex05.us.ad.irmc.com> <4DED3BFA.10500@candelatech.com> <80C343A5D090C24F9229C34CCD17287405F37ACA@usnjpar1blex05.us.ad.irmc.com> <4DEE6EA7.5080105@ll.mit.edu> Message-ID: <4E8A545B.2090401@candelatech.com> On 06/07/2011 11:32 AM, Jeff Mitchell wrote: > On 06/06/2011 04:49 PM, Grinberg, Leo wrote: >> No, this is a dedicated CentOS server. I have exactly the same server >> running xorp with no problems. > > Dedicated CentOS *what* server? > > FWIW, I've used it on 5.5 without problems, although it was significantly slower to start up than on same-spec'd Ubuntu and Gentoo VMs. Just to close this out: The problem was bad handling of some OSPF retransmit timers. I fixed this a few days ago and the test machine Iqor let me use started working much better.... Anyone using xorp 1.7 or higher and OSPF should upgrade to the latest github snapshot. I had unfortunately just released 1.8.4, so 1.8.5 will be out fairly soon... Thanks, Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com From sarada.ece08 at gmail.com Mon Oct 3 22:23:26 2011 From: sarada.ece08 at gmail.com (sarada lenka) Date: Tue, 4 Oct 2011 10:53:26 +0530 Subject: [Xorp-users] Functionality of virtual link is not working in ospf Message-ID: Hi, Xorp Version : 1.8.2 I have tested the below topology for virtual link in ospf Topology : H1 ----> R1 ----> R2 ----> R3 -----> H2 ---A0---- --------- A1 -------- ---A2--- Here, R1, which is connected to host(H1) is configured in area A0(0.0.0.0), other interface is configured in A1(0.0.0.1). Router R2 two interfaces are configured in A1(0.0.0.1) Router R3 which is connected to R2 is configured in area A1(0.0.0.1), other interface which is connected to host(H2) is configured in A2(0.0.0.2) virtual link has configured for R1 and R3. when i have executed the command "show route table ipv4 unicast final" in R1, it is not showing the R3 N/w connected to H2 Similarly when i have executed same command in R3, it is not showing the R1 N/w connected to H1 Regards, Sarada Devi.L -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.ICSI.Berkeley.EDU/pipermail/xorp-users/attachments/20111004/81b6e13a/attachment.html From greearb at candelatech.com Tue Oct 4 09:50:12 2011 From: greearb at candelatech.com (Ben Greear) Date: Tue, 04 Oct 2011 09:50:12 -0700 Subject: [Xorp-users] Functionality of virtual link is not working in ospf In-Reply-To: References: Message-ID: <4E8B3944.10402@candelatech.com> On 10/03/2011 10:23 PM, sarada lenka wrote: > Hi, > > Xorp Version : 1.8.2 > I have tested the below topology for virtual link in ospf > > Topology : > > H1 ----> R1 ----> R2 ----> R3 -----> H2 > ---A0---- --------- A1 -------- ---A2--- > > > Here, R1, which is connected to host(H1) is configured in area A0(0.0.0.0), other interface is configured in A1(0.0.0.1). > Router R2 two interfaces are configured in A1(0.0.0.1) > Router R3 which is connected to R2 is configured in area A1(0.0.0.1), other > interface which is connected to host(H2) is configured in A2(0.0.0.2) > > virtual link has configured for R1 and R3. > > when i have executed the command "show route table ipv4 unicast final" in R1, it is not showing the R3 N/w connected to H2 > Similarly when i have executed same command in R3, it is not showing the R1 N/w connected to H1 I haven't tested this particular setup. I did fix some other OSPF problems recently, so please see if the top-of-tree xorp from github works any better. If it does not, please post all of your xorp configs, and look through your logs for any obvious errors, and post the logs if you see something that looks suspicious. Tell us what OS you are using as well. Thanks, Ben > > Regards, > Sarada Devi.L > > > _______________________________________________ > Xorp-users mailing list > Xorp-users at xorp.org > http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-users -- Ben Greear Candela Technologies Inc http://www.candelatech.com From greearb at candelatech.com Tue Oct 4 11:21:48 2011 From: greearb at candelatech.com (Ben Greear) Date: Tue, 04 Oct 2011 11:21:48 -0700 Subject: [Xorp-users] xorp VLAN create/delete usage question Message-ID: <4E8B4EBC.7080102@candelatech.com> I just noticed that the vlan delete logic is messing up one of my own xorp configurations. Basically, I want to delete a vlan from the xorp config, but I don't want it deleted from the OS. I can easily enough add some hack like an environment variable to disable all vlan deletes, but I was wondering instead if I should be more clever and maybe have xorp only delete vlans that it created? Any opinions on this? Thanks, Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com From esj at cs.fiu.edu Tue Oct 4 11:47:48 2011 From: esj at cs.fiu.edu (Eric S. Johnson) Date: Tue, 04 Oct 2011 14:47:48 -0400 Subject: [Xorp-users] xorp VLAN create/delete usage question In-Reply-To: <4E8B4EBC.7080102@candelatech.com> References: <4E8B4EBC.7080102@candelatech.com> Message-ID: <20111004184748.C01E13680008@cheetah.cs.fiu.edu> greearb at candelatech.com said: > I just noticed that the vlan delete logic is messing up one of my > own xorp configurations. Basically, I want to delete a vlan from > the xorp config, but I don't want it deleted from the OS. > I can easily enough add some hack like an environment variable to > disable all vlan deletes, but I was wondering instead if I should be > more clever and maybe have xorp only delete vlans that it created? If you delete an interface that was previously created outside xotp it deletes it. So I would think vlan behavior should follow that paradigm, by default. So I would vote for a env variable based behavior change. E From greearb at candelatech.com Tue Oct 4 11:53:25 2011 From: greearb at candelatech.com (Ben Greear) Date: Tue, 04 Oct 2011 11:53:25 -0700 Subject: [Xorp-users] xorp VLAN create/delete usage question In-Reply-To: <20111004184748.C01E13680008@cheetah.cs.fiu.edu> References: <4E8B4EBC.7080102@candelatech.com> <20111004184748.C01E13680008@cheetah.cs.fiu.edu> Message-ID: <4E8B5625.5000401@candelatech.com> On 10/04/2011 11:47 AM, Eric S. Johnson wrote: > > greearb at candelatech.com said: >> I just noticed that the vlan delete logic is messing up one of my >> own xorp configurations. Basically, I want to delete a vlan from >> the xorp config, but I don't want it deleted from the OS. > >> I can easily enough add some hack like an environment variable to >> disable all vlan deletes, but I was wondering instead if I should be >> more clever and maybe have xorp only delete vlans that it created? > > If you delete an interface that was previously created outside xotp > it deletes it. It does now...question is whether it should? Xorp has no concept of delete interface v/s remove xorp configuration for interface. I was thinking that if xorp didn't create the VLAN itself, it should not delete the VLAN from the OS when removing xorp configuration for that interface. > > So I would think vlan behavior should follow that paradigm, by default. > So I would vote for a env variable based behavior change. That's easily enough done, but might not be flexible enough for all users... Perhaps a better way would be an explicit per-interface flag that users could set..but that is a lot code... Thanks, Ben > > E -- Ben Greear Candela Technologies Inc http://www.candelatech.com From jobhunts02 at aol.com Tue Oct 4 15:11:24 2011 From: jobhunts02 at aol.com (jobhunts02 at aol.com) Date: Tue, 4 Oct 2011 18:11:24 -0400 (EDT) Subject: [Xorp-users] (no subject) Message-ID: <8CE50F9B044A755-1FE4-BA2C@webmail-m021.sysops.aol.com> http://mustafaozsimseklerhoca.info/poster.php From greearb at candelatech.com Tue Oct 4 16:07:06 2011 From: greearb at candelatech.com (Ben Greear) Date: Tue, 04 Oct 2011 16:07:06 -0700 Subject: [Xorp-users] xorp VLAN create/delete usage question In-Reply-To: <4E8B5625.5000401@candelatech.com> References: <4E8B4EBC.7080102@candelatech.com> <20111004184748.C01E13680008@cheetah.cs.fiu.edu> <4E8B5625.5000401@candelatech.com> Message-ID: <4E8B919A.7030508@candelatech.com> On 10/04/2011 11:53 AM, Ben Greear wrote: > On 10/04/2011 11:47 AM, Eric S. Johnson wrote: >> >> greearb at candelatech.com said: >>> I just noticed that the vlan delete logic is messing up one of my >>> own xorp configurations. Basically, I want to delete a vlan from >>> the xorp config, but I don't want it deleted from the OS. >> >>> I can easily enough add some hack like an environment variable to >>> disable all vlan deletes, but I was wondering instead if I should be >>> more clever and maybe have xorp only delete vlans that it created? >> >> If you delete an interface that was previously created outside xotp >> it deletes it. > > It does now...question is whether it should? Xorp has no concept > of delete interface v/s remove xorp configuration for interface. > > I was thinking that if xorp didn't create the VLAN itself, it > should not delete the VLAN from the OS when removing xorp > configuration for that interface. I just committed code that will cause xorp to only delete from the OS VLANs that it created earlier. If anyone wants an env-variable to either force the VLANS to always be deleted, or never, I'll be happy to add it. Adding a new flag to the configuration to allow users per-interface control of vlan deletion is more effort than I feel like putting in, but patches are welcome. Thanks, Ben > >> >> So I would think vlan behavior should follow that paradigm, by default. >> So I would vote for a env variable based behavior change. > > That's easily enough done, but might not be flexible enough > for all users... > > Perhaps a better way would be an explicit per-interface flag that > users could set..but that is a lot code... > > Thanks, > Ben > > >> >> E > > -- Ben Greear Candela Technologies Inc http://www.candelatech.com From igorm at etf.rs Wed Oct 5 01:44:55 2011 From: igorm at etf.rs (=?ISO-8859-2?Q?Igor_Maravi=E6?=) Date: Wed, 5 Oct 2011 10:44:55 +0200 Subject: [Xorp-users] IEEE HPSR 2012 Message-ID: ********************************************************************************************************************* ********************************************************************************************************************* IEEE HPSR 2012 June 24-27 Belgrade Serbia http://www.ieee-hpsr.org Important Dates Full paper submission: January 20, 2012 Tutorial submission: January 20, 2012 Acceptance notifications: March 20, 2012 Camera-ready due: April 15, 2012 IEEE Conference on High Performance Switching and Routing 2012 will be held in Belgrade, the capital of Serbia. Belgrade is a lively metropolis that is a blend of different cultures whose representatives were visiting at various points of history. As such, it is very well suited to host the IEEE HPSR conference that gathers scientists and engineers from the entire world, after Heidelberg, Dallas, Kobe, Torino, Arizona, Hong Kong, Poznan, New York, Shanghai, Paris, Dallas (again), and Cartagena. Internet traffic keeps growing thanks to the access technologies whose speeds are increasing at a fast pace, and thanks to many non-profit and commercial efforts to bridge the digital-divide by providing all the advantages of the Internet to developing countries as well. At the same time, challenging applications are gaining popularity on the Internet such as multimedia and P2P applications. Multimedia applications require a lot of bandwidth and low delays, while the traffic patterns of P2P applications are impossible to predict. IPv4 addresses are being exhausted at this very moment. Data centers are moving a vast amount of content through thousands of switches. Also, routers and switches are being attacked, and energy available for them, as for everything else, is decreasing. All these circumstances put a higher burden than ever on switching and routing. So, there is a lot of work to be done, and IEEE HPSR 2012 will do its part by addressing the following topics: Architectures of high-performance switches and routers High-speed packet processors Address lookup algorithms Packet classification, scheduling and dropping Switching, bridging and routing protocols Multicasting P2P routing Routing in wireless, mobile and sensor networks Optical switching and routing Switching, bridging and routing in data centers and clouds Data placement and migration Multiprocessor networks Network management Pricing, accounting and charging QoS and scalability of switching, bridging and routing Traffic characterization and engineering Power-aware switching, bridging and routing protocols High-speed network security Paper submission guidelines: Submitted papers must be unpublished, and cannot be submitted elsewhere at the same time. Accepted papers should not exceed 6 pages in two-column IEEE Transactions style (www.ieee.org). Accepted papers longer than 6 pages will be charged $100 for each extra page. Papers cannot be longer than 8 pages. Papers should be submitted as PDF files through the EDAS system (http://edas.info). All submitted papers will be subject to three independent reviews. Awards: The Best Paper Award will be granted based on the reviewers? comments and scores that it received. Presentation of the paper will be scored by the audience, and will influence its final ranking. Note: In order to publish the paper at the IEEE HPSR 2012 proceedings and IEEE Xplore, the paper must be presented at the conference, and at least one of its authors should have a full registration. Non-refundable registration fees must be paid prior to uploading the final IEEE formatted, publication-ready, version of the paper. General Chairs: Aleksandra Smiljani? (Belgrade University, Serbia), Mounir Hamdi (Hong Kong University of Science and Technology, China) Technical Program Committee Chairs: Jonathan Chao (Polytechnic Institute of NYU, USA), Eiji Oki (University of Electro-Communications in Tokyo, Japan), Cyriel Minkenberg (IBM Research, Switzerland) Publicity Chairs: Andrea Bianco (Politecnico di Torino, Italy), Aleksandar Kolarov (Telcordia, USA), Dominique Verchere (Alcatel-Lucent, France) Registration and Finance Chair: Milan Bjelica (Belgrade University, Serbia) Publication Chair: Ljiljana Trajkovi? (Simon Fraser University, Canada) Local Arrangement Chairs: Zoran ?i?a, Nata?a Maksi?, and Marija Anti? (Belgrade University, Serbia) Tutorials Chair: Dejan Kosti? (EPFL, Switzerland), dejan.kostic at epfl.ch Panel Chair: Mitchell Gusat (IBM, Switzerland) Q&A: aleksandra at etf.rs, zoran.cica at etf.rs From JoeCoco at mecnet.net Wed Oct 5 05:03:51 2011 From: JoeCoco at mecnet.net (Joe Coco) Date: Wed, 5 Oct 2011 12:03:51 +0000 Subject: [Xorp-users] xorp VLAN create/delete usage question In-Reply-To: <4E8B919A.7030508@candelatech.com> References: <4E8B4EBC.7080102@candelatech.com> <20111004184748.C01E13680008@cheetah.cs.fiu.edu> <4E8B5625.5000401@candelatech.com> <4E8B919A.7030508@candelatech.com> Message-ID: <9513B0D809E3024A85A611E9B3726CA99C335C@MECnetExchange.mecnet.net> Hello, I kind of like the idea of either an ENV or a startup option. Leaning more towards the startup option. We're (considering) using Xorp in the following scenarios: A. Gui creates interfaces/vlans, and XORP just performs dynamic routing B. XORP does everything C. Initial interface configuration created at startup, XORP has control later on. I like the idea that based on a flag or switch, XORP would allow the adding, deleting, or modifying of any existing interface, or based on a different flag or switch XORP would either leave any interfaces configured alone on exit, or cleanup on exit and discard them. As to your initial concern, I don't completely understand why you would want to delete an interface from XORP's configuration but not have that interface modified in the OS. Could you explain that a little? -- Joe -----Original Message----- From: xorp-users-bounces at xorp.org [mailto:xorp-users-bounces at xorp.org] On Behalf Of Ben Greear Sent: Tuesday, October 04, 2011 7:07 PM To: Eric S. Johnson Cc: xorp-users at xorp.org Subject: Re: [Xorp-users] xorp VLAN create/delete usage question On 10/04/2011 11:53 AM, Ben Greear wrote: > On 10/04/2011 11:47 AM, Eric S. Johnson wrote: >> >> greearb at candelatech.com said: >>> I just noticed that the vlan delete logic is messing up one of my >>> own xorp configurations. Basically, I want to delete a vlan from >>> the xorp config, but I don't want it deleted from the OS. >> >>> I can easily enough add some hack like an environment variable to >>> disable all vlan deletes, but I was wondering instead if I should be >>> more clever and maybe have xorp only delete vlans that it created? >> >> If you delete an interface that was previously created outside xotp >> it deletes it. > > It does now...question is whether it should? Xorp has no concept of > delete interface v/s remove xorp configuration for interface. > > I was thinking that if xorp didn't create the VLAN itself, it should > not delete the VLAN from the OS when removing xorp configuration for > that interface. I just committed code that will cause xorp to only delete from the OS VLANs that it created earlier. If anyone wants an env-variable to either force the VLANS to always be deleted, or never, I'll be happy to add it. Adding a new flag to the configuration to allow users per-interface control of vlan deletion is more effort than I feel like putting in, but patches are welcome. Thanks, Ben > >> >> So I would think vlan behavior should follow that paradigm, by default. >> So I would vote for a env variable based behavior change. > > That's easily enough done, but might not be flexible enough for all > users... > > Perhaps a better way would be an explicit per-interface flag that > users could set..but that is a lot code... > > Thanks, > Ben > > >> >> E > > -- Ben Greear Candela Technologies Inc http://www.candelatech.com _______________________________________________ Xorp-users mailing list Xorp-users at xorp.org http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-users From greearb at candelatech.com Wed Oct 5 08:38:58 2011 From: greearb at candelatech.com (Ben Greear) Date: Wed, 05 Oct 2011 08:38:58 -0700 Subject: [Xorp-users] xorp VLAN create/delete usage question In-Reply-To: <9513B0D809E3024A85A611E9B3726CA99C335C@MECnetExchange.mecnet.net> References: <4E8B4EBC.7080102@candelatech.com> <20111004184748.C01E13680008@cheetah.cs.fiu.edu> <4E8B5625.5000401@candelatech.com> <4E8B919A.7030508@candelatech.com> <9513B0D809E3024A85A611E9B3726CA99C335C@MECnetExchange.mecnet.net> Message-ID: <4E8C7A12.2040300@candelatech.com> On 10/05/2011 05:03 AM, Joe Coco wrote: > Hello, > > I kind of like the idea of either an ENV or a startup option. Leaning more towards the startup option. > > We're (considering) using Xorp in the following scenarios: > > A. Gui creates interfaces/vlans, and XORP just performs dynamic routing > B. XORP does everything > C. Initial interface configuration created at startup, XORP has control later on. > > I like the idea that based on a flag or switch, XORP would allow the adding, deleting, or modifying of any existing interface, or based on a different flag or switch XORP > would either leave any interfaces configured alone on exit, or cleanup on exit and discard them. > > As to your initial concern, I don't completely understand why you would want to delete an interface from XORP's configuration but not have that interface modified in the OS. > > Could you explain that a little? I have an application that drives creation/deletion of interfaces, and sets up routing rules for them and such. I just need xorp to run the actual routing protocols. I need to be able to move interfaces in and out of xorp 'virtual routers' without them being deleted. I'll work on some env-var over-rides. Startup options are not easy because we do not directly start the xorp_fea process. It could be added to the xorp config, but that's a lot of work: I'd do it for hire, but I think not just for fun :) Thanks, Ben > > -- Joe > > -----Original Message----- > From: xorp-users-bounces at xorp.org [mailto:xorp-users-bounces at xorp.org] On Behalf Of Ben Greear > Sent: Tuesday, October 04, 2011 7:07 PM > To: Eric S. Johnson > Cc: xorp-users at xorp.org > Subject: Re: [Xorp-users] xorp VLAN create/delete usage question > > On 10/04/2011 11:53 AM, Ben Greear wrote: >> On 10/04/2011 11:47 AM, Eric S. Johnson wrote: >>> >>> greearb at candelatech.com said: >>>> I just noticed that the vlan delete logic is messing up one of my >>>> own xorp configurations. Basically, I want to delete a vlan from >>>> the xorp config, but I don't want it deleted from the OS. >>> >>>> I can easily enough add some hack like an environment variable to >>>> disable all vlan deletes, but I was wondering instead if I should be >>>> more clever and maybe have xorp only delete vlans that it created? >>> >>> If you delete an interface that was previously created outside xotp >>> it deletes it. >> >> It does now...question is whether it should? Xorp has no concept of >> delete interface v/s remove xorp configuration for interface. >> >> I was thinking that if xorp didn't create the VLAN itself, it should >> not delete the VLAN from the OS when removing xorp configuration for >> that interface. > > I just committed code that will cause xorp to only delete from the OS VLANs that it created earlier. > > If anyone wants an env-variable to either force the VLANS to always be deleted, or never, I'll be happy to add it. > > Adding a new flag to the configuration to allow users per-interface control of vlan deletion is more effort than I feel like putting in, but patches are welcome. > > Thanks, > Ben > >> >>> >>> So I would think vlan behavior should follow that paradigm, by default. >>> So I would vote for a env variable based behavior change. >> >> That's easily enough done, but might not be flexible enough for all >> users... >> >> Perhaps a better way would be an explicit per-interface flag that >> users could set..but that is a lot code... >> >> Thanks, >> Ben >> >> >>> >>> E >> >> > > > -- > Ben Greear > Candela Technologies Inc http://www.candelatech.com > > _______________________________________________ > Xorp-users mailing list > Xorp-users at xorp.org > http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-users -- Ben Greear Candela Technologies Inc http://www.candelatech.com From jobhunts02 at aol.com Sun Oct 16 05:04:25 2011 From: jobhunts02 at aol.com (jobhunts02 at aol.com) Date: Sun, 16 Oct 2011 08:04:25 -0400 (EDT) Subject: [Xorp-users] (no subject) Message-ID: <8CE5A12E1F9D49D-164-4BB89@webmail-d066.sysops.aol.com> http://www.metalurgicaoberto.com.ar/page.php?id=22 From JoeCoco at mecnet.net Tue Oct 18 06:41:13 2011 From: JoeCoco at mecnet.net (Joe Coco) Date: Tue, 18 Oct 2011 13:41:13 +0000 Subject: [Xorp-users] BGP policy bug (ipv6) Message-ID: <9513B0D809E3024A85A611E9B3726CA99C5A13@MECnetExchange.mecnet.net> Hello, Anyone ever take a look at this bug a little more or since March: http://www.xorp.org/bugzilla/show_bug.cgi?id=31 We really need to be able to apply per-peer policies on our BGP6 peers. n Joe -------------- next part -------------- An HTML attachment was scrubbed... URL: http://mailman.ICSI.Berkeley.EDU/pipermail/xorp-users/attachments/20111018/ce02a553/attachment.html From jobhunts02 at aol.com Sun Oct 30 21:34:04 2011 From: jobhunts02 at aol.com (jobhunts02 at aol.com) Date: Mon, 31 Oct 2011 00:34:04 -0400 (EDT) Subject: [Xorp-users] (no subject) Message-ID: <8CE659D74208CF2-1E78-106D09@webmail-d159.sysops.aol.com> http://hindustan247.com/detail.php?id=71