In other words, prplMesh is not listening on the ucc port.
This is the stdout content when starting prplMesh:
/opt/prplmesh/scripts/prplmesh_utils.sh: start
prplmesh_framework_init - starting local_bus and ieee1905_transport processes...
/opt/prplmesh/scripts/prplmesh_utils.sh: line 97: ebtables: not found
prplmesh_agent_start - start beerocks_agent process...
json_object_from_file: error opening file /tmp/share/logging.conf: No such file or directory
10:56:06 logger.cpp[144]: Configuration file does not exist
json_object_from_file: error opening file /tmp/share/logging.conf: No such file or directory
10:56:06 logger.cpp[144]: Configuration file does not exist
2020-06-14 10:56:06,470 DEBUG [default] [root@unknown-host] [int beerocks::bpl::cfg_get_all_prplmesh_wifi_interfaces(beerocks::bpl::BPL_WLAN_IFACE*, int*)] [prplmesh-1.4.0-86cbf8de/framework/platform/bpl/uci/cfg/bpl_cfg.cpp:393] cfg_get_all_prplmesh_wifi_interfaces: failed to get wifi interface for radio2 or radio2.hostap_iface doesn't exist
2020-06-14 10:56:06,471 DEBUG [default] [root@unknown-host] [int main(int, char**)] [prplmesh-1.4.0-86cbf8de/agent/src/beerocks/slave/beerocks_slave_main.cpp:513] radio0.hostap_iface=wlan0
2020-06-14 10:56:06,472 DEBUG [default] [root@unknown-host] [int main(int, char**)] [prplmesh-1.4.0-86cbf8de/agent/src/beerocks/slave/beerocks_slave_main.cpp:513] radio1.hostap_iface=wlan2
2020-06-14 10:56:06,473 DEBUG [default] [root@unknown-host] [static void beerocks::os_utils::kill_pid(const string&, const string&)] [prplmesh-1.4.0-86cbf8de/common/beerocks/bcl/source/beerocks_os_utils.cpp:94] kill_pid SIGTERM pid=9325
The kill_pid SIGTERM pid=9325 part seem to indicate another prplmesh process is still running when we call /opt/prplmesh/scripts/prplmesh_utils.sh: start.
When a process cannot be stopped with SIGTERM, beerocks::os_utils::kill_pid is supposed to wait for 15 seconds, then send it a SIGKILL. In this case however, it doesn't happen and we never reach the point where the slaves are started.
See for example this job: https://gitlab.com/prpl-foundation/prplMesh/-/jobs/604999730
It's also happening in the "nightly" tests from today (https://gitlab.com/prpl-foundation/prplMesh/-/jobs/606694852). The first tests were fine, then from
MAP-4.6.2_ETH_FH5GH
onward all the tests started failing like this:In other words, prplMesh is not listening on the ucc port.
This is the stdout content when starting prplMesh:
The
kill_pid SIGTERM pid=9325
part seem to indicate another prplmesh process is still running when we call/opt/prplmesh/scripts/prplmesh_utils.sh: start
.When a process cannot be stopped with SIGTERM,
beerocks::os_utils::kill_pid
is supposed to wait for 15 seconds, then send it a SIGKILL. In this case however, it doesn't happen and we never reach the point where the slaves are started.