PS. The same effect can be achieved by binding a TCP service other than Redis to the port Kamailio expects Redis to be on (e.g. 6379).
On 11/11/2014 08:38 PM, Alex Balashov wrote:
If I kill my Redis server and boot Kamailio with an ndb_redis configuration to connect to Redis via TCP, I get an infinite loop of:
Nov 11 20:37:00 centosity6 /usr/local/sbin/kamailio[1892]: ERROR: <core> [io_wait.h:1048]: io_wait_loop_epoll(): epoll_wait(29, 0x7fe025ea06f8, 0, 5000): Invalid argument [22] Nov 11 20:37:00 centosity6 /usr/local/sbin/kamailio[1892]: ERROR: <core> [io_wait.h:1048]: io_wait_loop_epoll(): epoll_wait(29, 0x7fe025ea06f8, 0, 5000): Invalid argument [22] Nov 11 20:37:00 centosity6 /usr/local/sbin/kamailio[1892]: ERROR: <core> [io_wait.h:1048]: io_wait_loop_epoll(): epoll_wait(29, 0x7fe025ea06f8, 0, 5000): Invalid argument [22] Nov 11 20:37:00 centosity6 /usr/local/sbin/kamailio[1892]: ERROR: <core> [io_wait.h:1048]: io_wait_loop_epoll(): epoll_wait(29, 0x7fe025ea06f8, 0, 5000): Invalid argument [22] Nov 11 20:37:00 centosity6 /usr/local/sbin/kamailio[1892]: ERROR: <core> [io_wait.h:1048]: io_wait_loop_epoll(): epoll_wait(29, 0x7fe025ea06f8, 0, 5000): Invalid argument [22] Nov 11 20:37:00 centosity6 /usr/local/sbin/kamailio[1892]: ERROR: <core> [io_wait.h:1048]: io_wait_loop_epoll(): epoll_wait(29, 0x7fe025ea06f8, 0, 5000): Invalid argument [22] Nov 11 20:37:00 centosity6 /usr/local/sbin/kamailio[1892]: ERROR: <core> [io_wait.h:1048]: io_wait_loop_epoll(): epoll_wait(29, 0x7fe025ea06f8, 0, 5000): Invalid argument [22] Nov 11 20:37:00 centosity6 /usr/local/sbin/kamailio[1892]: ERROR: <core> [io_wait.h:1048]: io_wait_loop_epoll(): epoll_wait(29, 0x7fe025ea06f8, 0, 5000): Invalid argument [22] Nov 11 20:37:00 centosity6 /usr/local/sbin/kamailio[1892]: ERROR: <core> [io_wait.h:1048]: io_wait_loop_epoll(): epoll_wait(29, 0x7fe025ea06f8, 0, 5000): Invalid argument [22] [...] <ad infinitum/nauseum/full filesystem>
-- Alex