Bug #6352
openttcn3-bts-test[-latest] is broken since Feb 1st / docker and multicast
100%
Description
ttcn3-bts-test[-latest] is failing for a few days already:
https://jenkins.osmocom.org/jenkins/view/TTCN3/job/ttcn3-bts-test/2293/
https://jenkins.osmocom.org/jenkins/view/TTCN3/job/ttcn3-bts-test-latest/1953/
The console log (https://jenkins.osmocom.org/jenkins/view/TTCN3/job/ttcn3-bts-test/2293/console) tells us about problems with the virtphy:
+ docker kill jenkins-ttcn3-bts-test-2293-virtphy Error response from daemon: Cannot kill container: jenkins-ttcn3-bts-test-2293-virtphy: No such container: jenkins-ttcn3-bts-test-2293-virtphy
Here is the related logging (https://jenkins.osmocom.org/jenkins/view/TTCN3/job/ttcn3-bts-test/2293/artifact/logs/bts/osmo-bts.log):
0: starting: osmo-bts-virtual -c /data/osmo-bts.gen.cfg ((*)) | / \ OsmoBTS 20240201061853485 DLGLOBAL NOTICE Unimplemented bts_model_phy_instance_set_defaults (main.c:156) 20240201061853485 DLGLOBAL NOTICE Unimplemented bts_model_phy_instance_set_defaults (main.c:156) 20240201061853485 DLGLOBAL NOTICE Unimplemented bts_model_phy_instance_set_defaults (main.c:156) 20240201061853485 DLGLOBAL NOTICE Unimplemented bts_model_phy_instance_set_defaults (main.c:156) 20240201061853486 DLGLOBAL NOTICE Setting up GSMTAP Um forwarding '(null)->'172.18.29.10:4729' (main.c:363) 20240201061853486 DLCTRL NOTICE CTRL at 0.0.0.0 4238 (control_if.c:1014) 20240201061853487 DL1C NOTICE Unimplemented bts_model_ctrl_cmds_install (bts_model.c:222) 20240201061853487 DLGLOBAL NOTICE Available via telnet 0.0.0.0 4241 (telnet_interface.c:88) 20240201061853487 DPCU INFO Started listening on PCU socket (PCU IF v12): /data/unix/pcu_sock (pcu_sock.c:1237) 20240201061853487 DOSMUX INFO Osmux socket listening on 172.18.29.20:1984 (osmux.c:352) 20240201061853487 DABIS NOTICE A-bis connection establishment to BSC (127.0.0.11) in progress... (abis.c:161) 20240201061853487 DLINP NOTICE enabling ipaccess BTS mode, OML connecting to 127.0.0.11:3002 (ipaccess.c:1098) 20240201061853487 DL1C INFO phy0: PHY link state change shutdown -> connecting (phy_link.c:58) Failed to join to mcast goup: No such device Unable to create VirtualUm multicast socket: No such device 20240201061853487 DL1C INFO phy0: PHY link state change connecting -> shutdown (phy_link.c:58) unable to open PHY link(s) 0: stopped pid 8 with status 2
The virtphy container (https://jenkins.osmocom.org/jenkins/view/TTCN3/job/ttcn3-bts-test/2293/artifact/logs/virtphy/virtphy.log) fails to start:
Thu Feb 1 06:18:53 2024 DVIRPHY virtphy.c:248 Virtual physical layer starting up... Failed to join to mcast goup: No such device Unable to create VirtualUm multicast socket: No such device Segmentation fault (core dumped)
Updated by fixeria 3 months ago
Interestingly enough, this does not affect CentOS:
https://jenkins.osmocom.org/jenkins/view/TTCN3-centos/job/TTCN3-centos-bts-test/
https://jenkins.osmocom.org/jenkins/view/TTCN3-centos/job/TTCN3-centos-bts-test-latest/
Both testsuites look normal.
Updated by fixeria 3 months ago
- Status changed from In Progress to Feedback
https://gerrit.osmocom.org/c/docker-playground/+/35908 ttcn3-bts-test: do not start osmo-bts-virtual / virtphy containers
Updated by fixeria 3 months ago
fixeria wrote in #note-3:
https://gerrit.osmocom.org/c/docker-playground/+/35908 ttcn3-bts-test: do not start osmo-bts-virtual / virtphy containers
This did not fix the problem because I forgot to comment out some more lines:
https://gerrit.osmocom.org/c/docker-playground/+/35941 fixup: ttcn3-bts-test: do not start osmo-bts-virtual / virtphy containers [NEW]
Updated by fixeria 3 months ago
laforge FYI, I tried running the virtphy based testcases on the build machine (container) you allocated to me, and surprisingly they all passed. So the multicast is actually working between docker containers.
Could it be that there were some changes on the build machine(s) which are executing TTCN-3 testcases, disabling the multicast support somehow?
Updated by laforge 2 months ago
- Subject changed from ttcn3-bts-test[-latest] is broken since Feb 1st to ttcn3-bts-test[-latest] is broken since Feb 1st / docker and multicast
- Assignee changed from fixeria to osmith
- Priority changed from Normal to Low
assinging to osmith at a low priority. I don't know why the docker setup of our jenkins slaves (docker in lxc containers) would not support multicast while it works on fixeria's (docker in lxc container) on the very same physical machine.
Updated by fixeria about 2 months ago
- Status changed from Feedback to Resolved
- % Done changed from 0 to 100
Applied in changeset core-testing-infra:docker-playground|8c6ebc8c764fb4d7170372112649a87e3c440d83.
Updated by fixeria about 2 months ago
fixeria [not really, it's actually Redmine] wrote in #note-7:
Applied in changeset core-testing-infra:docker-playground|8c6ebc8c764fb4d7170372112649a87e3c440d83.
Uh, "Closes: OS#xxxx" works blazing fast... ~26 days after the patch got merged.
We still want to get virtphy based tests running in Docker, so let's keep this ticket open.