Thanks Jason. I don't think the issue is with IGMP anymore...all my systems are running v3, and the switches are also forced to v3. It is NOT a problem with Ganglia because we can't get any other multicast traffic to pass, either. I'm going to try a few more experiments before moving to another Cisco switch.
Steve Gilbert Unix Systems Administrator [EMAIL PROTECTED] -----Original Message----- From: Jason A. Smith [mailto:[EMAIL PROTECTED] Sent: Wednesday, September 15, 2004 10:43 AM To: Peter Schmid Cc: Steve Gilbert; Ganglia General Subject: RE: [Ganglia-general] Any gmond issues on Opteron? You might be interested to know that RedHat has re-opened the igmp bug, and has posted a kernel patch that adds the force igmp version option to sysctl. It also appears to fix the version identification in /proc. They said this will probably be added to the RHEL3 4th quarter update in December: https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=125570 ~Jason On Thu, 2004-09-09 at 14:20, Peter Schmid wrote: > That's all right... the source is "supposed" to fail over to V2 if the > switches do not support V3... but apparently it does not do so well at > that ;) > > On Thu, 9 Sep 2004, Steve Gilbert wrote: > > > We just looked at the RedHat kernel source...sure enough...it checks > > to see if it's v1, and it not, it just blindly puts "v2" in > > /proc/net/igmp, even though it's really talking v3. Style. > > > > So all my machines are really talking v3...I'll force v3 on the > > switches now and see if that helps. > > > > > > Steve Gilbert > > Unix Systems Administrator > > [EMAIL PROTECTED] > > > > -----Original Message----- > > From: Peter Schmid [mailto:[EMAIL PROTECTED] > > Sent: Thursday, September 09, 2004 11:00 AM > > To: Steve Gilbert > > Cc: [email protected] > > Subject: RE: [Ganglia-general] Any gmond issues on Opteron? > > > > A quick way to tell... is take RH EL 2.1 or RH 8.0 machine and add > > it to the opteron cluster....Both those OS's do not do IGMP v3 (with > > stock kernel). And gmond should see the rest of the Opteorn machines. > > > > pete. > > > > On Thu, 9 Sep 2004, Steve Gilbert wrote: > > > > > Whoops...I think I spoke too soon...maybe this is the issue after > > all... > > > > > > All of our /proc/net/igmp files report v2, but I just did a > > > tcpdump, and there is a lot of v3 on there. > > > > > > So I guess part of the bug in RHEL 3 is that is reports v2, but > > > actually talks v3? > > > > > > > > > Steve Gilbert > > > Unix Systems Administrator > > > [EMAIL PROTECTED] > > > > > > -----Original Message----- > > > From: Peter Schmid [mailto:[EMAIL PROTECTED] > > > Sent: Thursday, September 09, 2004 10:20 AM > > > To: Steve Gilbert > > > Cc: [email protected] > > > Subject: Re: [Ganglia-general] Any gmond issues on Opteron? > > > > > > Steve, > > > > > > There are some older posts from me and a few others who ran into this. > > > > > It's not the opteron per se.. but if you are running a OS that > > > uses IGMP v3... there is a bug in RHEL 3 that will not fail over > > > to IGMP > > v2. > > > > > > Have your network people sniff the traffic to see if Ganglia is > > > using IGMP v3.. and either they can enable it..or in my case.. I > > > have a IGMP > > > v2 node that is mute.. .and I use it to gather the infro and feed > > > to gmetad for the IGMP v3 nodes. > > > > > > HTH, > > > > > > Pete. > > > > > > On Thu, 9 Sep 2004, Steve Gilbert wrote: > > > > > > > Are there any known bugs or other issues with running gmond on > > > > AMD Opteron based systems? > > > > > > > > I just built gmond 2.5.6 from source on a small Opteron > > > > cluster...no > > > > > > errors or warnings or anything during the complile/install process. > > > > Gmond seems to be running fine, but none of my nodes can see > > > > each > > > other. > > > > gstat --all on each machine just shows itself...the others > > > > aren't even > > > > > > > showing up in the Dead list. > > > > > > > > Even though my network team can't find anything wrong, I suspect > > > > that I might be having another issue with multicast...but > > > > thought I should see if there are any Opteron issues. > > > > > > > > Thanks. > > > > > > > > Steve Gilbert > > > > Unix Systems Administrator > > > > [EMAIL PROTECTED] > > > > > > > > > > > > ------------------------------------------------------- > > > > This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of > > > > 170 Project Admins to receive an Apple iPod Mini FREE for your > > > > judgement > > > > > > on who ports your project to Linux PPC the best. Sponsored by IBM. > > > > Deadline: Sept. 13. Go here: http://sf.net/ppc_contest.php > > > > _______________________________________________ > > > > Ganglia-general mailing list > > > > [email protected] > > > > https://lists.sourceforge.net/lists/listinfo/ganglia-general > > > > > > > > > > -- > > > Peter Schmid > > > Technical Director > > > Logic Technology Inc. > > > GE Global Research Center > > > ITMS Engineering Systems Group > > > (518) 387-6903 > > > > > > > -- > > Peter Schmid > > Technical Director > > Logic Technology Inc. > > GE Global Research Center > > ITMS Engineering Systems Group > > (518) 387-6903 > > -- /------------------------------------------------------------------\ | Jason A. Smith Email: [EMAIL PROTECTED] | | Atlas Computing Facility, Bldg. 510M Phone: (631)344-4226 | | Brookhaven National Lab, P.O. Box 5000 Fax: (631)344-7616 | | Upton, NY 11973-5000 | \------------------------------------------------------------------/
