[technique] Problèmes de déconnexion machine Myrys

[Laurent GUERBY]

On Wed, 2012-07-18 at 00:09 +0200, Solarus wrote:
> Le 17/07/2012 19:59, Laurent GUERBY a écrit :
> > On Tue, 2012-07-17 at 19:14 +0200, Solarus wrote:
> >
> > Salut Solarus,
> >
> > Depuis le dimanche 15 juillet 21h-23h nous avons 3 VM et une machine
> > physique qui ont des pertes de ping regulieres d'apres nagios, la
> > machine physique est proxmox.ultrawaves.fr (celle qu'on surveille avec
> > nagios).
> >
> > Sincèrement,
> >
> > Laurent
> >
> Merci beaucoup, je vais suivre ça avec Tuxfanou.

Bonjour,

J'ai un debut de piste, depuis h3 91.224.149.254 00:01:2e:2b:d2:07 je
ping proxmox 91.224.149.139 00:22:4d:54:a2:da et je surveille
au tcpdump -eni eth0.3131 ether host 00:22:4d:54:a2:da

08:19:41.359442 00:01:2e:2b:d2:07 > 00:22:4d:54:a2:da, ethertype IPv4 (0x0800), length 98: 91.224.149.254 > 91.224.149.139: ICMP echo request, id 25871, seq 212, length 64
08:19:41.359624 00:22:4d:54:a2:da > 00:01:2e:2b:d2:07, ethertype IPv4 (0x0800), length 98: 91.224.149.139 > 91.224.149.254: ICMP echo reply, id 25871, seq 212, length 64
08:19:42.361082 00:01:2e:2b:d2:07 > 00:22:4d:54:a2:da, ethertype IPv4 (0x0800), length 98: 91.224.149.254 > 91.224.149.139: ICMP echo request, id 25871, seq 213, length 64
08:19:42.361419 00:22:4d:54:a2:da > 00:01:2e:2b:d2:07, ethertype IPv4 (0x0800), length 98: 91.224.149.139 > 91.224.149.254: ICMP echo reply, id 25871, seq 213, length 64
08:19:43.363068 00:01:2e:2b:d2:07 > 00:22:4d:54:a2:da, ethertype IPv4 (0x0800), length 98: 91.224.149.254 > 91.224.149.139: ICMP echo request, id 25871, seq 214, length 64
08:19:43.363362 00:22:4d:54:a2:da > 00:01:2e:2b:d2:07, ethertype IPv4 (0x0800), length 98: 91.224.149.139 > 91.224.149.254: ICMP echo reply, id 25871, seq 214, length 64

sequence normale de ping, et puis :

08:19:44.362078 00:01:2e:2b:d2:07 > 00:22:4d:54:a2:da, ethertype IPv4 (0x0800), length 98: 91.224.149.254 > 91.224.149.139: ICMP echo request, id 25871, seq 215, length 64
08:19:44.362388 00:22:4d:54:a2:da > 00:01:2e:2b:d2:07, ethertype IPv4 (0x0800), length 98: 91.224.149.139 > 91.224.149.254: ICMP echo reply, id 25871, seq 215, length 64
08:19:44.615843 00:22:4d:54:a2:da > 01:00:5e:00:00:01, ethertype IPv4 (0x0800), length 60: 0.0.0.0 > 224.0.0.1: igmp query v2
08:19:44.615898 00:22:4d:54:a2:da > 33:33:00:00:00:01, ethertype IPv6 (0x86dd), length 86: 100:0:600:0:e0db:100:: > ff02::1: HBH ICMP6, multicast listener querymax resp delay: 1000 addr: ::, length 24
08:19:44.632976 00:22:4d:54:a2:da > 01:00:5e:00:00:01, ethertype IPv4 (0x0800), length 60: 0.0.0.0 > 224.0.0.1: igmp query v2

ces 3 paquets IGMP ont l'air de venir de proxmox et a partir de la il n'y a plus de reply au ping :

08:19:45.364234 00:01:2e:2b:d2:07 > 00:22:4d:54:a2:da, ethertype IPv4 (0x0800), length 98: 91.224.149.254 > 91.224.149.139: ICMP echo request, id 25871, seq 216, length 64
08:19:46.364322 00:01:2e:2b:d2:07 > 00:22:4d:54:a2:da, ethertype IPv4 (0x0800), length 98: 91.224.149.254 > 91.224.149.139: ICMP echo request, id 25871, seq 217, length 64

... jusqu'a ce que h3 fasse un arp request whos-has :

08:20:29.448354 00:01:2e:2b:d2:07 > 00:22:4d:54:a2:da, ethertype IPv4 (0x0800), length 98: 91.224.149.254 > 91.224.149.139: ICMP echo request, id 25871, seq 260, length 64
08:20:30.244192 00:01:2e:2b:d2:07 > 00:22:4d:54:a2:da, ethertype ARP (0x0806), length 42: Request who-has 91.224.149.139 tell 91.224.149.254, length 28
08:20:30.448334 00:01:2e:2b:d2:07 > 00:22:4d:54:a2:da, ethertype IPv4 (0x0800), length 98: 91.224.149.254 > 91.224.149.139: ICMP echo request, id 25871, seq 261, length 64
08:20:30.573971 00:01:2e:2b:d2:07 > 00:22:4d:54:a2:da, ethertype IPv4 (0x0800), length 98: 46.165.195.139 > 91.224.149.139: ICMP echo request, id 61974, seq 1, length 64

Avec la VM tetalab (91.224.149.202 aa:00:00:a1:5e:f8) qui a le meme
soucis de coupure que proxmox j'ai observé exactement le meme
comportement, tcpdump plus verbose cette fois ci :

08:36:24.652596 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 667, length 64
08:36:24.653425 aa:00:00:a1:5e:f8 > 00:01:2e:2b:d2:07, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 5295, offset 0, flags [none], proto ICMP (1), length 84)
    91.224.149.202 > 91.224.149.254: ICMP echo reply, id 27649, seq 667, length 64
08:36:25.653877 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 668, length 64
08:36:25.654556 aa:00:00:a1:5e:f8 > 00:01:2e:2b:d2:07, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 5296, offset 0, flags [none], proto ICMP (1), length 84)
    91.224.149.202 > 91.224.149.254: ICMP echo reply, id 27649, seq 668, length 64
08:36:26.420014 aa:00:00:a1:5e:f8 > 01:00:5e:4d:7c:d5, ethertype IPv4 (0x0800), length 60: (tos 0xc0, ttl 1, id 0, offset 0, flags [DF], proto IGMP (2), length 32, options (RA))
    91.224.149.202 > 239.77.124.213: igmp v2 report 239.77.124.213
08:36:26.444109 aa:00:00:a1:5e:f8 > 01:00:5e:4d:7c:d5, ethertype IPv4 (0x0800), length 60: (tos 0xc0, ttl 1, id 0, offset 0, flags [DF], proto IGMP (2), length 32, options (RA))
    91.224.149.202 > 239.77.124.213: igmp v2 report 239.77.124.213
08:36:26.655342 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 669, length 64
08:36:27.662663 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 670, length 64
...
08:36:48.814402 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 691, length 64
08:36:49.687752 aa:00:00:a1:5e:f8 > 01:00:5e:4d:7c:d5, ethertype IPv4 (0x0800), length 60: (tos 0xc0, ttl 1, id 0, offset 0, flags [DF], proto IGMP (2), length 32, options (RA))
    91.224.149.202 > 239.77.124.213: igmp v2 report 239.77.124.213
08:36:49.716793 aa:00:00:a1:5e:f8 > 01:00:5e:4d:7c:d5, ethertype IPv4 (0x0800), length 60: (tos 0xc0, ttl 1, id 0, offset 0, flags [DF], proto IGMP (2), length 32, options (RA))
    91.224.149.202 > 239.77.124.213: igmp v2 report 239.77.124.213
08:36:49.823568 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 692, length 64
...
08:37:04.928293 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype ARP (0x0806), length 42: Ethernet (len 6), IPv4 (len 4), Request who-has 91.224.149.202 tell 91.224.149.254, length 28
08:37:04.934803 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 707, length 64
08:37:05.929278 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype ARP (0x0806), length 42: Ethernet (len 6), IPv4 (len 4), Request who-has 91.224.149.202 tell 91.224.149.254, length 28
08:37:05.942617 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 708, length 64
08:37:06.929417 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype ARP (0x0806), length 42: Ethernet (len 6), IPv4 (len 4), Request who-has 91.224.149.202 tell 91.224.149.254, length 28
08:37:06.944313 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 709, length 64
08:37:07.949185 aa:00:00:a1:5e:f8 > 00:01:2e:2b:d2:07, ethertype ARP (0x0806), length 60: Ethernet (len 6), IPv4 (len 4), Reply 91.224.149.202 is-at aa:00:00:a1:5e:f8, length 46
08:37:07.949234 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 710, length 64
08:37:07.949680 aa:00:00:a1:5e:f8 > 00:01:2e:2b:d2:07, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 5297, offset 0, flags [none], proto ICMP (1), length 84)
    91.224.149.202 > 91.224.149.254: ICMP echo reply, id 27649, seq 710, length 64
08:37:08.946215 00:01:2e:2b:d2:07 > aa:00:00:a1:5e:f8, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto ICMP (1), length 84)
    91.224.149.254 > 91.224.149.202: ICMP echo request, id 27649, seq 711, length 64
08:37:08.947007 aa:00:00:a1:5e:f8 > 00:01:2e:2b:d2:07, ethertype IPv4 (0x0800), length 98: (tos 0x0, ttl 64, id 5298, offset 0, flags [none], proto ICMP (1), length 84)
    91.224.149.202 > 91.224.149.254: ICMP echo reply, id 27649, seq 711, length 64

La aussi il y a une requete IGMP qui declenche la coupure du ping jusqu'a une serie d'arp request who-has. 

En cherchant 239.77.124.213 je tombe sur :

https://wiki.jenkins-ci.org/display/JENKINS/Auto-discovering+Jenkins+on+the+network
Auto-discovering Jenkins on the network
(...)
Jenkins listens on UDP port 33848. You can either send an UDP broadcast
packet (targeted to 255.255.255.255) to this port, or you can send an
UDP multicast packet (targeted to 239.77.124.213) to this port
(...)

https://wiki.jenkins-ci.org/display/JENKINS/Meet+Jenkins
Jenkins is an award-winning application that monitors executions of
repeated jobs, such as building a software project or jobs run by cron.

La VM tetalab auquelle j'ai acces a bien un jenkins qui tourne, peut-etre c'est
le cas sur les autres machines avec le probleme ? Peut-etre plusieurs
Jenkins sur le meme LAN necessite quelques reglages ?

Merci a tous,

Laurent