[FUG-BR] Problema com Watchdog de NICs

[Enio Marconcini]

pessoal, percebi recentemente que, num servidor rodando 7.2-Release (isso
mesmo, meu primeiro teste com freebsd) vez ou outra os usuários ligam
reclamando que perdeu conexão com o sistema

no começo eu até pensei que era os AP´s que estavam travando, mas depois de
olhar o messages eu notei essas informações

Jan 13 12:08:39 glaurung kernel: vge0: link state changed to UP
Jan 13 12:36:56 glaurung kernel: vge1: watchdog timeout
Jan 13 12:36:56 glaurung kernel: vge1: link state changed to DOWN
Jan 13 12:36:57 glaurung kernel: vge1: link state changed to UP
Jan 13 12:48:33 glaurung kernel: vge0: watchdog timeout
Jan 13 12:48:33 glaurung kernel: vge0: link state changed to DOWN
Jan 13 12:48:36 glaurung kernel: vge0: link state changed to UP
Jan 13 13:31:12 glaurung kernel: arp: unknown hardware address format
(0x4500)

Jan 13 15:46:37 glaurung kernel: vge1: watchdog timeout
Jan 13 15:46:37 glaurung kernel: vge1: link state changed to DOWN
Jan 13 15:46:39 glaurung kernel: vge1: link state changed to UP
Jan 13 15:46:47 glaurung kernel: vge1: watchdog timeout
*Jan 13 15:46:47 glaurung kernel: miibus3: unknown CICADA PHY model 0*
Jan 13 15:46:47 glaurung kernel: vge1: link state changed to DOWN
Jan 13 15:46:49 glaurung kernel: vge1: link state changed to UP
Jan 13 15:47:05 glaurung kernel: vge1: watchdog timeout
Jan 13 15:47:05 glaurung kernel: vge1: link state changed to DOWN
Jan 13 15:47:07 glaurung kernel: vge1: link state changed to UP
Jan 13 15:47:15 glaurung kernel: vge1: watchdog timeout
Jan 13 15:47:15 glaurung kernel: vge1: link state changed to DOWN
Jan 13 15:47:17 glaurung kernel: vge1: link state changed to UP
*Jan 13 15:48:12 glaurung kernel: vge0: failed to idle MII autopoll
Jan 13 15:48:12 glaurung kernel: vge2: failed to idle MII autopoll*


dentre essas linhas, as que eu marquei com negritão me são totalmente
estranhas e procurando na net por CICADA PHY eu nao encontrei muita
explicação


nos momentos que a rede cai, eu dei um netstat pra ver se tava sem buffers
de memória de rede mas estava aparentemente normal

as informações das minhas placas são essas:

*glaurung# ifconfig vge0
vge0: flags=8847<UP,BROADCAST,DEBUG,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu
1500
        options=1b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING>
        ether 00:0c:42:1a:56:14
        inet 10.10.10.17 netmask 0xfffffff0 broadcast 10.10.10.31
        inet 10.10.3.33 netmask 0xffffffe0 broadcast 10.10.3.63
        inet 10.10.3.65 netmask 0xffffffe0 broadcast 10.10.3.95
        inet 10.10.3.97 netmask 0xffffffe0 broadcast 10.10.3.127
        inet 10.10.4.33 netmask 0xffffffe0 broadcast 10.10.4.63
        media: Ethernet autoselect (100baseTX <full-duplex>)
        status: active*

*glaurung# ifconfig vge1
vge1: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
        options=1b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING>
        ether 00:0c:42:1a:56:15
        inet 10.10.10.1 netmask 0xfffffff0 broadcast 10.10.10.15
        inet 10.10.3.129 netmask 0xffffffe0 broadcast 10.10.3.159
        media: Ethernet autoselect (100baseTX <full-duplex>)
        status: active**
*


eu to achando estranho essas mensagens do kernel sobre Watchdog

alguém poderia me esclarecer?

abraços

-- 
ENIO RODRIGO MARCONCINI
gtalk: eniorm em gmail.com
skype: eniorm
msn: /dev/null

.: FreeBSD -:- OpenBSD -:-Slackware Linux :.