So everything was great for a couple of hours, until the traffic started to ramp up (233 logged in via free group).
Jan 8 19:04:05 grase01 coova-chilli[1381]: net.c: 780: 11 (Resource temporarily unavailable) could not read packet
Jan 8 19:04:05 grase01 coova-chilli[1381]: net.c: 806: 11 (Resource temporarily unavailable) net_read_eth(fd=5, len=65497, mtu=1500) == -1
Jan 8 19:04:05 grase01 coova-chilli[1381]: net.c: 780: 11 (Resource temporarily unavailable) could not read packet
Jan 8 19:04:05 grase01 coova-chilli[1381]: net.c: 806: 11 (Resource temporarily unavailable) net_read_eth(fd=5, len=65497, mtu=1500) == -1
Jan 8 19:04:05 grase01 coova-chilli[1381]: net.c: 780: 11 (Resource temporarily unavailable) could not read packet
Jan 8 19:04:05 grase01 coova-chilli[1381]: net.c: 806: 11 (Resource temporarily unavailable) net_read_eth(fd=5, len=65497, mtu=1500) == -1
Jan 8 19:04:05 grase01 coova-chilli[1381]: net.c: 780: 11 (Resource temporarily unavailable) could not read packet
Jan 8 19:04:05 grase01 coova-chilli[1381]: net.c: 806: 11 (Resource temporarily unavailable) net_read_eth(fd=5, len=65497, mtu=1500) == -1
It doesn't look like linux exhausted its sockets, I saw under 350 active sockets (netstat -an | wc -l).
According to the logs the above error was appearing for maybe 5 minutes, but still curious if there is another limit we need to increase.
There is talk about the "enable large limits" option when compiling coova-chilli - do you know if this is was enabled for the custom compile of coova-chilli?
Thanks,