On 29/04/18 18:58, Stuart Henderson wrote: > On 2018-04-29, Stuart Longland <[email protected]> wrote: >> The rack has 5 servers, a ARM-based PC and the switch, all of which run >> from a pair of 12V 105Ah AGM batteries, charged from mains power and >> solar. Switch is a Linksys LGS326-AU. No other devices plugged into >> this switch have connectivity issues. >> >> The port the industrial PC is connected to is a plain access port with >> no VLAN tagging, trunking or other funny stuff (although all of the >> above get used elsewhere in the network). >> >> When the link drops out, there's nothing in `dmesg`. If I tether my >> phone and hit the machine via SSH, I find it is unable to ping anything >> on the internal network via IPv4 or IPv6, or vice versa. > > What does "ifconfig rl0" show, both normally and when this happens? > > How about "netstat -nI rl0"?
Not sure about netstat; but ifconfig rl0 isn't any different during and outside of inaccessibility events. Right now, I'm getting this: > # ifconfig rl0 > > rl0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500 > lladdr 00:d0:c9:e0:f4:75 > index 1 priority 0 llprio 3 > media: Ethernet autoselect (100baseTX full-duplex) > status: active > inet 172.31.249.254 netmask 0xffffff00 broadcast 172.31.249.255 > inet6 fe80::2d0:c9ff:fee0:f475%rl0 prefixlen 64 scopeid 0x1 > inet6 2001:44b8:21ac:70f9::fe prefixlen 64> # netstat -nI rl0 > Name Mtu Network Address Ipkts Ierrs Opkts Oerrs > Colls > rl0 1500 <Link> 00:d0:c9:e0:f4:75 13402102 0 16803928 0 > 0 > rl0 1500 172.31.249/ 172.31.249.254 13402102 0 16803928 0 > 0 > rl0 1500 fe80::%rl0/ fe80::2d0:c9ff:fe 13402102 0 16803928 0 > 0 > rl0 1500 2001:44b8:2 2001:44b8:21ac:70 13402102 0 16803928 0 > 0 I'll try both commands again when I get problems. > Is anything logged on the switch? This model has things like broadcast > storm control, I'm wondering if that might have triggered or if it shows > anything else useful. I'll have a closer look, there were some messages about the link going up and down, but it seems I missed turning the SNTP client on (settings in two places) so the log timestamps are years off. I've reset the logs and will see how we go. > Can you try a different cable, can you try a different switch port? Unfortunately all ports are full. I did replace the cable though. >> Is there some sort of debugging flag I can turn on in the kernel to log >> more detail about what's going on with rl0 when the loss of connectivity >> is being experienced? > > I don't see any extra debugging that can be enabled for rl. > No problems, well, I'll keep an eye on the switch and see what it tells me. -- Stuart Longland (aka Redhatter, VK4MSL) I haven't lost my mind... ...it's backed up on a tape somewhere.
