Hi, Ray, thanks for the tips... Yes, it seems that XORP is happier with a complete configuration file. However, I still don't get multicast routing to work end-2-end with xorp.
Thanks! On Thu, Dec 8, 2011 at 12:16 PM, Ray Soucy <[email protected]> wrote: > XORP is production quality in terms of its functionality. The XORP > shell doesn't work very well, and is "touchy" when trying to > interactively configure it. > > We script the generation of the "config.boot" file, and boot XORP > directly off the pre-written configuration, rather than trying to > configure it through the XORP shell. > > We have not had any major stability or performance issues as a result. > > To my knowledge, there are no other open source projects that > correctly support multicast. The closest is the pimd project for > Quagga, but it was very limited in its functionality last time I > checked (only supports SSM for example). > > The syntax of the XORP CLI is similar to what you would see from > Juniper. The majority of network engineers can intuitively make use > of it; bugs aside. I don't think it needs to be re-written to mirror > Cisco. It just needs to be cleaned up. Perhaps by someone who would > find value in such functionality. > > On Thu, Dec 8, 2011 at 2:51 PM, frank hu <[email protected]> wrote: > > > > Thanks Ben for the kind help! > > > > I did use the debuggers to look into the core dumps. However, I expected > > xorp to be reasonably stable and such core dumps to happen only > > occasionally. From what I can see, the features XORP multicast routing > > engine offers are pretty nice -- not too complex but good enough > for even a > > lot enterprise users. However, the software suite doesn't seem mature > enough > > to be bundled in industrial grade network devices. > > > > Just some comment, if XORP also targets outside the academia community, > its > > overall configuration mechanism should be designed to be intuitive for > Cisco > > IOS users for the simple reason that most router admins are much familiar > > with Cisco IOS CLIs. For example, why a multicast routing admin (of XORP) > > needs to manually add register_vif interface? The software suite needs > to do > > that under the hood, instead of having the user do it. > > > > > > I think that my assessment of XORP multicast routing will stop here. > > > > Also thanks a lot to Ray and everyone else that helped me through the > issues > > that I ran into -- I appreciate your time and your in-depth expertise in > > XORP. > > > > Frank > > > > > > On Thu, Dec 8, 2011 at 11:11 AM, Ben Greear <[email protected]> > wrote: > >> > >> On 12/08/2011 11:08 AM, frank hu wrote: > >>> > >>> > >>> Hi, > >>> I still need more help. > >>> I rebooted my x86 box with xorp and restarted the xorp configuration > from > >>> scratch. I configured PIM over the interface "vlan300" and IGMP over > >>> "vlan800" and > >>> committed the configurations. It went well. Then I tried to add > >>> plumbing/mfea4 and also the PIM register interface "register_vif". > >>> Unfortunately it failed at > >>> this step. > >> > >> > >> Look for core files and errors in the xorp logs. If you find cores, use > >> gdb to get > >> a backtrace (compile with debugging symbols if you haven't already). > >> > >> Use 'ulimit -c unlimited' before starting xorp_rtrmgr if it is crashing > >> and > >> not dumping core files. > >> > >> Thanks, > >> Ben > >> > >> > >>> So the critical process "xorp_rtrmgr" was gone, along with "xorp_pim" > and > >>> "xorp_igmp". > >>> It seems that a configuration change triggered the above to happen. > >> > >> > >> > >> -- > >> Ben Greear <[email protected]> > >> Candela Technologies Inc http://www.candelatech.com > >> > > > > > > -- > Ray Soucy > > Epic Communications Specialist > > Phone: +1 (207) 561-3526 > > Networkmaine, a Unit of the University of Maine System > http://www.networkmaine.net/ >
_______________________________________________ Xorp-hackers mailing list [email protected] http://mailman.ICSI.Berkeley.EDU/mailman/listinfo/xorp-hackers
