Kernel oooops lorsque beaucoup de traffic réseau

Bonsoir,

J’ai un curieux problème sur l’un de mes serveurs.

Basiquement, il se manifeste en donnant la commande suivante :
mv /home/somedirectory /mnt/nfsdisk/somedirectory & disown

La copie commence puis quelques minutes plus tard s’arrête sans raison.
/mnt/nfsdisk est un dossier monté en NFS sur IPv4 sans aucun paramètre spécial.

Le serveur est Dell PowerEdge 850
Le kernel est (Debian lenny): Linux serveur-1 2.6.26-2-amd64 #1 SMP Fri Mar 27 04:02:59 UTC 2009 x86_64 GNU/Linux
Le load: load average: 2.29, 2.05, 1.94
Et dans le syslog:

[quote]Apr 23 21:14:32 serveur-1 kernel: [91057.108437] ------------[ cut here ]------------
Apr 23 21:14:32 serveur-1 kernel: [91057.108470] WARNING: at net/core/dev.c:1451 skb_gso_segment+0x88/0x1af()
Apr 23 21:14:32 serveur-1 kernel: [91057.108503] Modules linked in: xt_multiport appletalk crc32c libcrc32c nfnetlink_queue nfnetlink nf_conntrack_ipv4 ipt_REJECT xt_tcpudp iptable_filter ip_tables xt_iprange xt_state nf_conntrack xt_mark xt_NFQUEUE x_tables ipmi_devintf ipmi_si ipmi_msghandler ib_iser rdma_cm ib_cm iw_cm ib_sa ib_mad ib_core ib_addr iscsi_tcp libiscsi scsi_transport_iscsi nfsd auth_rpcgss exportfs nfs lockd nfs_acl sunrpc bonding ipv6 ext2 mbcache loop snd_pcm snd_timer snd soundcore snd_page_alloc serio_raw psmouse pcspkr iTCO_wdt i2c_i801 i2c_core rng_core button i3000_edac edac_core shpchp pci_hotplug evdev dcdbas joydev xfs sha256_generic aes_x86_64 aes_generic cbc dm_crypt crypto_blkcipher dm_mirror dm_log dm_snapshot dm_mod ide_floppy ide_cd_mod cdrom ata_generic libata dock usbhid hid ff_memless sg sd_mod piix siimage ide_pci_generic ide_core arcmsr ehci_hcd scsi_mod uhci_hcd tg3 thermal processor fan thermal_sys
Apr 23 21:14:32 serveur-1 kernel: [91057.110280] Pid: 0, comm: swapper Tainted: G W 2.6.26-2-amd64 #1
Apr 23 21:14:32 serveur-1 kernel: [91057.110320]
Apr 23 21:14:32 serveur-1 kernel: [91057.110321] Call Trace:
Apr 23 21:14:32 serveur-1 kernel: [91057.110373] [] warn_on_slowpath+0x51/0x7a
Apr 23 21:14:32 serveur-1 kernel: [91057.110461] [] :nf_conntrack:__nf_ct_refresh_acct+0x108/0x15c
Apr 23 21:14:32 serveur-1 kernel: [91057.110554] [] :ipv6:fib6_lookup_1+0x95/0x109
Apr 23 21:14:32 serveur-1 kernel: [91057.110628] [] :ipv6:find_match+0x104/0x160
Apr 23 21:14:32 serveur-1 kernel: [91057.110711] [] :ipv6:ip6_pol_route+0x1d0/0x296
Apr 23 21:14:32 serveur-1 kernel: [91057.110770] [] skb_gso_segment+0x88/0x1af
Apr 23 21:14:32 serveur-1 kernel: [91057.110823] [] dev_hard_start_xmit+0x17f/0x24e
Apr 23 21:14:32 serveur-1 kernel: [91057.110878] [] dev_queue_xmit+0x1f0/0x2b2
Apr 23 21:14:32 serveur-1 kernel: [91057.110949] [] :ipv6:ip6_output+0xc11/0xc37
Apr 23 21:14:32 serveur-1 kernel: [91057.110998] [] fib_rules_lookup+0x7f/0xb3
Apr 23 21:14:32 serveur-1 kernel: [91057.111083] [] :ipv6:ip6_xmit+0x35c/0x44e
Apr 23 21:14:32 serveur-1 kernel: [91057.111179] [] :ipv6:inet6_csk_xmit+0x219/0x235
Apr 23 21:14:32 serveur-1 kernel: [91057.111275] [] tcp_transmit_skb+0x739/0x776
Apr 23 21:14:32 serveur-1 kernel: [91057.111332] [] tso_fragment+0xef/0x13b
Apr 23 21:14:32 serveur-1 kernel: [91057.111388] [] __tcp_push_pending_frames+0x6ba/0x78d
Apr 23 21:14:32 serveur-1 kernel: [91057.111454] [] tcp_rcv_established+0x80f/0x8b6
Apr 23 21:14:32 serveur-1 kernel: [91057.111534] [] :ipv6:tcp_v6_do_rcv+0x626/0x8ff
Apr 23 21:14:32 serveur-1 kernel: [91057.111588] [] tcp_rcv_established+0x5dc/0x8b6
Apr 23 21:14:32 serveur-1 kernel: [91057.111669] [] :ipv6:ip6_pol_route_input+0x0/0x11
Apr 23 21:14:32 serveur-1 kernel: [91057.111736] [] :ipv6:fib6_rule_action+0x79/0x196
Apr 23 21:14:32 serveur-1 kernel: [91057.111778] [] skb_checksum+0x4d/0x285
Apr 23 21:14:32 serveur-1 kernel: [91057.111823] [] __inet6_lookup_established+0xed/0x258
Apr 23 21:14:32 serveur-1 kernel: [91057.111884] [] :ipv6:tcp_v6_rcv+0x638/0x685
Apr 23 21:14:32 serveur-1 kernel: [91057.111941] [] :ipv6:ip6_input_finish+0x1d5/0x34c
Apr 23 21:14:32 serveur-1 kernel: [91057.111994] [] :ipv6:ipv6_rcv+0x32d/0x38d
Apr 23 21:14:32 serveur-1 kernel: [91057.112044] [] :tg3:tg3_poll+0x641/0x8f1
Apr 23 21:14:32 serveur-1 kernel: [91057.112107] [] net_rx_action+0xab/0x1da
Apr 23 21:14:32 serveur-1 kernel: [91057.112149] [] __do_softirq+0x5c/0xd1
Apr 23 21:14:32 serveur-1 kernel: [91057.112176] [] ack_apic_level+0x53/0xd8
Apr 23 21:14:32 serveur-1 kernel: [91057.112212] [] call_softirq+0x1c/0x28
Apr 23 21:14:32 serveur-1 kernel: [91057.112243] [] do_softirq+0x3c/0x81
Apr 23 21:14:32 serveur-1 kernel: [91057.112271] [] irq_exit+0x3f/0x83
Apr 23 21:14:32 serveur-1 kernel: [91057.112300] [] do_IRQ+0xb9/0xd9
Apr 23 21:14:32 serveur-1 kernel: [91057.112328] [] mwait_idle+0x0/0x4d
Apr 23 21:14:32 serveur-1 kernel: [91057.112357] [] ret_from_intr+0x0/0x19
Apr 23 21:14:32 serveur-1 kernel: [91057.112385] [] mwait_idle+0x41/0x4d
Apr 23 21:14:32 serveur-1 kernel: [91057.112449] [] cpu_idle+0x89/0xb3
Apr 23 21:14:32 serveur-1 kernel: [91057.112520]
Apr 23 21:14:32 serveur-1 kernel: [91057.112540] —[ end trace d73e16009c0227c9 ]—
[/quote]

A la suite d’une recherche sur Google, il semblerait que ce soit un bug du kernel lié à IPv6.
J’ai effectivement des adresses IPv6 fournies par Sixxs.

Est-ce que quelqu’un a une idée que faire pour éviter ceci ?

Merci de votre aide.

Ça se voit que c’est ipv6, c’est dans ces fonctions que ça plante. Rjaoute un fichier

/etc/modprobe.d/vireipv6

contenant

Merci pour ta réponse.
Idéalement j’aimerais garder IPv6…
Là je l’ai désactivé et rebooté, on va voir si ça change quelque chose.

Je comprends pas pourquoi le module IPv6 fait planter un NFS en IPv4, y-a-t-il une explication rationnelle ?

Juste pour info, même le ping en IPv6 renvoie des erreurs:

[quote]PING ipv6.google.com(tx-in-x68.google.com) 56 data bytes
64 bytes from tx-in-x68.google.com: icmp_seq=1 ttl=47 time=359 ms
64 bytes from tx-in-x68.google.com: icmp_seq=2 ttl=47 time=326 ms
64 bytes from tx-in-x68.google.com: icmp_seq=3 ttl=47 time=376 ms
64 bytes from tx-in-x68.google.com: icmp_seq=4 ttl=47 time=346 ms
64 bytes from tx-in-x68.google.com: icmp_seq=6 ttl=47 time=329 ms
64 bytes from tx-in-x68.google.com: icmp_seq=7 ttl=47 time=344 ms
64 bytes from tx-in-x68.google.com: icmp_seq=8 ttl=47 time=361 ms
64 bytes from tx-in-x68.google.com: icmp_seq=9 ttl=47 time=397 ms
64 bytes from tx-in-x68.google.com: icmp_seq=10 ttl=47 time=328 ms
ipv6.google.com ping statistics —
10 packets transmitted, 9 received, 10% packet loss, time 9038ms
rtt min/avg/max/mdev = 326.321/352.242/397.559/22.691 ms[/quote]

EDIT:
Cela ne semble pas avoir désactivé IPv6:

[quote]bond0 Link encap:Ethernet HWaddr 00:12:3f:2a:96:f4
inet addr:10.10.10.30 Bcast:10.10.10.255 Mask:255.255.255.0
inet6 addr: 2001:41e0:ff5f:0:212:3fff:fe2a:96f4/64 Scope:Global
inet6 addr: fe80::212:3fff:fe2a:96f4/64 Scope:Link
inet6 addr: 2001:41e0:ff5f:0:223:4cdf:0:1/64 Scope:Global
UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
RX packets:128787 errors:0 dropped:0 overruns:0 frame:0
TX packets:644846 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:23660729 (22.5 MiB) TX bytes:939539609 (896.0 MiB)
[/quote]