All connectivity dies every 24-48 hours on ReadyNAS Pro 6
I have an unusual problem with my ReadyNAS Pro 6 (4.2.22), which has only started happening the last couple of weeks. About every 24-48 hours the ethernet connection dies completely. The box looks like it's still alive and running, but no longer shows up in RAIDar and can't be pinged.
Checking in /var/log/syslog, I see the following unusual event --
Feb 2 16:10:28 machinename avahi-daemon[3111]: Received invalid packet. Feb 2 16:42:29 megalithica kernel: UDP: bad checksum. From 208.67.220.220:53 to 192.168.0.201:47330 ulen 37 Feb 2 23:59:17 machinename chronyd[2887]: Can't synchronise: no reachable sources Feb 2 23:59:19 machinename kernel: eth0: network connection down Feb 2 23:59:19 machinename ifplugd(eth0)[2753]: Link beat lost. Feb 2 23:59:20 machinename ifplugd(eth0)[2753]: Executing '/etc/ifplugd/ifplugd.action eth0 down'. Feb 2 23:59:20 machinename avahi-daemon[3111]: Withdrawing address record for 192.168.0.201 on eth0. Feb 2 23:59:20 machinename avahi-daemon[3111]: Leaving mDNS multicast group on interface eth0.IPv4 with address 192.168.0.201. Feb 2 23:59:20 machinename avahi-daemon[3111]: Interface eth0.IPv4 no longer relevant for mDNS. Feb 2 23:59:21 machinename ifplugd(eth0)[2753]: client: No udhcpc found running; none killed.
Looks like it happens right after avahi-daemon receives an invalid packet from OpenDNS. Some time after that it looks like the connection goes down. 192.168.0.201 is the ReadyNAS IP address.
I run rsync to synchronize files between a couple of machines, so I have been noticing this happening on a regular basis recently. Just in case, I'm going to switch out the ethernet cable and see if it helps. But this has me stumped, as I've been running without problems for quite some time.
Re: All connectivity dies every 24-48 hours on ReadyNAS Pro
Do you see anything about a 'fault' in the logs? Usually when the network stack dies, there is a GPF (general protection fault). Other users with GPF that see similar issue: viewtopic.php?f=24&t=67938
Re: All connectivity dies every 24-48 hours on ReadyNAS Pro
Checked the logs, don't see any GPF. I've now shut down DNLA services and turned off UDP forwarding for rsync connections (since it shouldn't be needed anyway). Haven't seen it go down for a couple of days now. I'll report back here next time something happens. Still not sure of the cause though.
Re: All connectivity dies every 24-48 hours on ReadyNAS Pro
Looks like there's still issues. Lost connectivity overnight, and the front panel wouldn't even respond (other than to completely power-off and reboot). Found masses of junk in syslog. Example --
Feb 10 08:03:29 machinename kernel: >WARNImsg+0g bug 43D D5DD r5D04449 flcomm: rnted: P Trace: Feb 10 08:03:29 machinename kernel: 50 Feb 10 08:03:29 machinename kernel: [<mon+0x8b/0x/0x70 Feb 10 08:03:29 machinename kernel: [<0 Feb 10 08:03:29 machinename kernel: [<fffffffx20 Feb 10 08:03:29 machinename kernel: [<0x30 Feb 10 08:03:29 machinename kernel: [0x33/0x40 Feb 10 08:03:29 machinename kernel: [<fffffx8bd/0xb5073/0x90 Feb 10 08:03:29 machinename kernel: <4150 Feb 10 08:03:29 machinename kernel: [<ffffffff884> [<ffffffff880fffffff8803> [<ffffffff88050 Feb 10 08:03:29 machinename kernel: [<ffff170 Feb 10 08:03:29 machinename kernel: [<f90 Feb 10 08:03:29 machinename kernel: [<fffff0x0/0x5 Feb 10 08:03:29 machinename kernel: ---[ end trace 400afb3762881696 ]--- Feb 10 08:03:29 machinename kernel: ---------t net/xb50(pied D0443DvnxD040 Feb 10 08:03:29 machinename kernel: Modul Tainte W 2. [<ffffffffffffff880c0 Feb 10 08:03:29 machinename kernel: [<ff<ffffffff880fffff88002bce>]fffffff885ffffffff884> [<ffffffff884fb18d<4> [<fffff [<ffffffffffffff882c2763>] ? ff880ae251>] do90d4>] ? t Feb 10 08:03:29 machinename kernel: ------------[ cut htcp.c:ardwa3D seqt D549 Feb 10 08:03:29 machinename kernel: <4s linked .37.6.RNx884fb18d>] 32b>] warffffff880bd034459>] warce>] ? apic_tim4531>] ? bb253>] ? ff884fb18d>] tcp_recfff88518ae3884b Feb 10 08:03:29 machinename kernel: ------------[ cut here ]----tcp_rece: To443D rfl 0 Feb 10 08:03:29 machinename kernel: Mod li: nv6gpio 2.6.37_64.2.4 #? tcp_recvslowpath_dc0>] ? polrn_slowpath__timer_interrupaw_spin_loock_sock_n] tcp_recvmsg+0x8bd/0>] inet_r Feb 10 08:03:29 machinename kernel: [<ead+0xd1/0x110 <DIED HERE>
Similar messages in kern.log. Looks like connectivity was lost a few hours earlier. Here's netatalk.log --
Feb 10 01:48:18.562871 afpd[28607] {socket.c:157} (E:AFPDaemon): read: Connection reset by peer message repeated 6 times Feb 10 01:48:18.562937 afpd[28607] {dsi_stream.c:421} (E:DSI): dsi_stream_read: len:-1, Connection reset by peer Feb 10 01:48:18.562963 afpd[28607] {dsi_stream.c:250} (N:DSI): dsi_disconnect: entering disconnected state
Really bizarre stuff, no idea what is causing the machine to go down every few days. Any ideas on where else I can investigate?
