Re: [OT] munin
Disabled ipv6 both through /etc/modprobe.d/aliases and through /proc, and it didn't make a difference. Connection still refused on port 4949 when telnetting to localhost. I'm beginning to wonder if it is a missing or interfering package. --b On Mon, Mar 5, 2012 at 1:30 AM, Wolfgang Karall li...@karall-edv.at wrote: On Sun, Mar 04, 2012 at 06:44:47PM -0500, Brad Alexander wrote: Yes. Non-working: echo -e 'list\nquit' | nc 192.168.4.9 4949 (UNKNOWN) [192.168.4.9] 4949 (munin) : Connection refused Hmm, since Bob suggest a bug in recent versions (I don't know, not running wheezy or sid on any munin-node installations at the moment), could you try disabling IPv6 and try again? And if this is really a bug with munin-node not working with IPv6 enabled this needs to be reported to the BTS. Cheers Wolfgang -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/cakmzw+zsmy3hxhl1lcxj+d_x5vhftn-dp28rb8s552hoy__...@mail.gmail.com
Re: [OT] munin
On Mon, Mar 5, 2012 at 2:32 PM, Brad Alexander stor...@gmail.com wrote: On Mon, Mar 5, 2012 at 1:30 AM, Wolfgang Karall li...@karall-edv.at wrote: On Sun, Mar 04, 2012 at 06:44:47PM -0500, Brad Alexander wrote: Yes. Non-working: echo -e 'list\nquit' | nc 192.168.4.9 4949 (UNKNOWN) [192.168.4.9] 4949 (munin) : Connection refused Hmm, since Bob suggest a bug in recent versions (I don't know, not running wheezy or sid on any munin-node installations at the moment), could you try disabling IPv6 and try again? And if this is really a bug with munin-node not working with IPv6 enabled this needs to be reported to the BTS. Disabled ipv6 both through /etc/modprobe.d/aliases and through /proc, and it didn't make a difference. Connection still refused on port 4949 when telnetting to localhost. I'm beginning to wonder if it is a missing or interfering package. On sid, ipv6 is compiled in to the kernel so it can only be disabled via ipv6.disable=1 on the linux line in grub.cfg. If you've compiled your own kernel with ipv6 as a module, have you tried disabling ipv6 with install ipv6 /bin/true in /etc/modprobe.d/ipv6.conf and rebuilt the initrd after running depmod -a? Is cat /proc/sys/net/ipv6/bindv6only the same on working and non-working nodes? This setting restricts the ipv6 socket to ipv6 (or not) and had created problems in the past for some apps. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/CAOdo=sznpu4+biw2v-jak4ttgqmzrpnv6xui83hctjbc0vv...@mail.gmail.com
Re: [OT] munin
Hello, why's this OT? :) On Sat, Mar 03, 2012 at 11:01:07PM -0500, Brad Alexander wrote: this is on both the working and non-working nodes. My network is not running ipv6. In fact, net-pf-10 is commented out on all the hosts that I have checked. IIRC, to actually disable IPv6 you need to actually uncomment the line, replacing the ipv6 with off, like alias net-pf-10 off Can anyone tell me why these 5 nodes will not connect to the server and refuses connections to localhost and ::1? What does grep ^allow /etc/munin/munin-node.conf produce on the machines? Cheers Wolfgang -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120304084825.ga16...@lenny.spiney.org
Re: [OT] munin
On Sun, Mar 4, 2012 at 3:48 AM, Wolfgang Karall lists+debian-u...@karall-edv.at wrote: Hello, why's this OT? :) Just because there are certain people on the list who are more draconian about what is on and off topic. To them, there is a difference between apps running *on* Debian and Debian itself. I figured I'd play it safe. :) On Sat, Mar 03, 2012 at 11:01:07PM -0500, Brad Alexander wrote: this is on both the working and non-working nodes. My network is not running ipv6. In fact, net-pf-10 is commented out on all the hosts that I have checked. IIRC, to actually disable IPv6 you need to actually uncomment the line, replacing the ipv6 with off, like alias net-pf-10 off Ah. I figured if the net-pf-10 was commented out, it was taken out of play. Can anyone tell me why these 5 nodes will not connect to the server and refuses connections to localhost and ::1? What does grep ^allow /etc/munin/munin-node.conf produce on the machines? The only one not commented out is allow ^192\.168\.4\.63$ Where 192.168.4.63 is the munin server. Note that with the exception of the hostname in the host_name directive, the config file is identical on all nodes, working and non-working. Puppet is doing a variable substitution for the hostname and pushing the file out to all nodes. The treatment is identical everywhere, which is what is confusing me why these 5 don't work. --b Cheers Wolfgang -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120304084825.ga16...@lenny.spiney.org -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/cakmzw+zzroet9sxuyttdps4vktgl+gn8tpqd1zgwf4cfbqa...@mail.gmail.com
Re: [OT] munin
On Sun, Mar 04, 2012 at 07:11:49AM -0500, Brad Alexander wrote: alias net-pf-10 off Ah. I figured if the net-pf-10 was commented out, it was taken out of play. No, you actually have to turn it off, commented out means use the defaults, i.e. IPv6 is available. Not that it matters I guess. The netstat output for many servers shows only TCP6/UDP6, but they still work on IPv4 as well (there are 2 ways to create the sockets, I can't remember the details), e.g. apache: lenny:~# netstat -tulpen | grep apache tcp6 0 0 :::80:::* LISTEN 0 2062901 7794/apache2 tcp6 0 0 :::443 :::* LISTEN 0 2062905 7794/apache2 The only one not commented out is allow ^192\.168\.4\.63$ Where 192.168.4.63 is the munin server. And from the munin server, what does echo -e 'list\nquit' | nc address according to munin.conf 4949 yield for the working/non-working hosts, any difference? Cheers Wolfgang PS: no need to CC me, I'm subscribed to the list. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120304205247.ga9...@lenny.spiney.org
Re: [OT] munin
Brad Alexander wrote: Can anyone tell me why these 5 nodes will not connect to the server and refuses connections to localhost and ::1? For what it is worth I have exactly the same problem on Sid. Works okay on Squeeze. Broken on Sid. Sid the munin-node doesn't bind to an IPv4 accessible address. This used to work fine in Sid up through around the end of September 2011 but then was broken by some upgrade. I note that it was broken then but forgot to go back and look as to why or what broke it. Your note here reminds me of this and tells me that I should go look to see why. All of my production systems run Stable and so hasn't been on my radar. Out of sight and out of mind. But definitely needs to be investigated and resolved before Wheezy releases. (That should put this clearly on-topic for debian-user. smile) Check the version of munin-node and dependencies of it on all of your systems and you will find it correlated to one of the versions. Bob signature.asc Description: Digital signature
Re: [OT] munin
On Sun, Mar 4, 2012 at 3:52 PM, Wolfgang Karall lists+debian-u...@karall-edv.at wrote: On Sun, Mar 04, 2012 at 07:11:49AM -0500, Brad Alexander wrote: alias net-pf-10 off Ah. I figured if the net-pf-10 was commented out, it was taken out of play. No, you actually have to turn it off, commented out means use the defaults, i.e. IPv6 is available. Not that it matters I guess. The netstat output for many servers shows only TCP6/UDP6, but they still work on IPv4 as well (there are 2 ways to create the sockets, I can't remember the details), e.g. apache: lenny:~# netstat -tulpen | grep apache tcp6 0 0 :::80 :::* LISTEN 0 2062901 7794/apache2 tcp6 0 0 :::443 :::* LISTEN 0 2062905 7794/apache2 The only one not commented out is allow ^192\.168\.4\.63$ Where 192.168.4.63 is the munin server. And from the munin server, what does echo -e 'list\nquit' | nc address according to munin.conf 4949 yield for the working/non-working hosts, any difference? Yes. Non-working: echo -e 'list\nquit' | nc 192.168.4.9 4949 (UNKNOWN) [192.168.4.9] 4949 (munin) : Connection refused And working: echo -e 'list\nquit' | nc 192.168.4.11 4949 # munin node at farragut cpu df df_inode entropy exim_mailqueue exim_mailstats forks fw_packets http_loadtime if_err_eth0 if_eth0 interrupts iostat iostat_ios irqstats load ntp_kernel_err ntp_kernel_pll_freq ntp_kernel_pll_off ntp_offset open_files open_inodes proc_pri processes swap threads uptime vmstat -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/cakmzw+btjgckms49kkelhvev2xzzghnora7me5qefmz0bdj...@mail.gmail.com
Re: [OT] munin
On Sun, Mar 4, 2012 at 4:20 PM, Bob Proulx b...@proulx.com wrote: Brad Alexander wrote: Can anyone tell me why these 5 nodes will not connect to the server and refuses connections to localhost and ::1? For what it is worth I have exactly the same problem on Sid. Works okay on Squeeze. Broken on Sid. Sid the munin-node doesn't bind to an IPv4 accessible address. My problem is that there is no rhyme or reason to this. Statistically, it seems to be across the board. I have 14 machines that I am trying to get set up. Of those, 7 are physical machines and 8 are OpenVZ containers. 9 are working, 5 are not. Here is what I am looking at statistically: Working: lenny: 1 squeeze: 3 wheezy: 1 sid: 4 Non-working: wheezy: 2 sid: 3 The versions in wheezy and sid are the same (1.4.6-3), the version in squeeze is 1.4.5-3 and the version in lenny is 1.2.6-10~lenny2. So since the problem seems to be so widespread, I'm guessing it has to be something with the way the machine is configured...I have looked at a lot of the usual suspects (logs, firewall, etc) but have come up empty. --b This used to work fine in Sid up through around the end of September 2011 but then was broken by some upgrade. I note that it was broken then but forgot to go back and look as to why or what broke it. Your note here reminds me of this and tells me that I should go look to see why. All of my production systems run Stable and so hasn't been on my radar. Out of sight and out of mind. But definitely needs to be investigated and resolved before Wheezy releases. (That should put this clearly on-topic for debian-user. smile) Check the version of munin-node and dependencies of it on all of your systems and you will find it correlated to one of the versions. Bob -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/cakmzw+zzg9faklghvxfs1cfd25q_kpoypanpmzco7unp+tz...@mail.gmail.com
Re: [OT] munin
On Sun, Mar 04, 2012 at 06:44:47PM -0500, Brad Alexander wrote: Yes. Non-working: echo -e 'list\nquit' | nc 192.168.4.9 4949 (UNKNOWN) [192.168.4.9] 4949 (munin) : Connection refused Hmm, since Bob suggest a bug in recent versions (I don't know, not running wheezy or sid on any munin-node installations at the moment), could you try disabling IPv6 and try again? And if this is really a bug with munin-node not working with IPv6 enabled this needs to be reported to the BTS. Cheers Wolfgang -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20120305063010.ga4...@lenny.spiney.org
[OT] munin
Hey all, I installed munin on my network, and wrote a puppet module to push out the munin-node.conf to each machine. Everything seems to work fine, and 9 machines are reporting in and generating graphs, but 5 are not. The configs are identical with the exception of the hostname. Here's the rub. Of the 5 machines, 4 are openvz containers, and one is a physical machine. 3 are running sid and 2 are running wheezy. (the 9 that are reporting are a mix of lenny, squeeze, wheezy and sid, and are both physical and openvz containers.) When I telnet to localhost 4949, it gives me connection refused on all 5 servers, but munin-node is running on all 6. Now here is something interesting. On all of the nodes I have checked, it appears to be listening on the ipv6 address: netstat -pant | grep 4949 tcp6 0 0 :::4949 :::* LISTEN 27585/munin-node this is on both the working and non-working nodes. My network is not running ipv6. In fact, net-pf-10 is commented out on all the hosts that I have checked. Here is a snippet from my /var/log/munin/munin-node.log. This was a /etc/init.d/munin-node restart, a telnet localhost 4949 (nothing written to the logs) and a telnet ::1 4949 2012/03/03-22:57:05 Server closing! Process Backgrounded 2012/03/03-22:57:06 Munin::Node::Server (type Net::Server::Fork) starting! pid(13974) Using default listen value of 128 Binding to TCP port 4949 on host * 2012/03/03-22:57:57 CONNECT TCP Peer: ::1:60551 Local: ::1:4949 2012/03/03-22:57:57 [14328] Denying connection from: ::1 Can anyone tell me why these 5 nodes will not connect to the server and refuses connections to localhost and ::1? Thanks, --b -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/cakmzw+amq8yzekavoc0hvb77882tjzaka2y4hnn190bzxyq...@mail.gmail.com