[fuel-migrate] sshd filed to start after first reboot of fuel node

[Ilya Bumarskov]

Steps to reproduce:

Run the fuel-migrate script and check logs of fuel node via RDP.

Observed behaviour:
OpenSSH server daemon failed to start after reboot master node. As a result, fuel-migrate script can't migrate database.

sshd.service logs:
[root@nailgun ~]# journalctl -u sshd.service
-- Logs begin at Fri 2017-07-21 10:09:58 UTC, end at Fri 2017-07-21 10:27:01 UTC. --
Jul 21 10:10:01 nailgun.test.domain.local systemd[1]: Installed new job sshd.service/start as 104
-- Logs begin at Fri 2017-07-21 10:09:58 UTC, end at Fri 2017-07-21 10:27:01 UTC. --
Jul 21 10:10:01 nailgun.test.domain.local systemd[1]: Installed new job sshd.service/start as 104
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: About to execute: /usr/sbin/sshd -D $OPTIONS
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: Forked /usr/sbin/sshd as 690
-- Logs begin at Fri 2017-07-21 10:09:58 UTC, end at Fri 2017-07-21 10:27:01 UTC. --
Jul 21 10:10:01 nailgun.test.domain.local systemd[1]: Installed new job sshd.service/start as 104
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: About to execute: /usr/sbin/sshd -D $OPTIONS
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: Forked /usr/sbin/sshd as 690
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: sshd.service changed dead -> start
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: Starting OpenSSH server daemon...
Jul 21 10:10:02 nailgun.test.domain.local systemd[690]: Executing: /usr/sbin/sshd -D
Jul 21 10:10:02 nailgun.test.domain.local sshd[690]: Could not load host key: /etc/ssh/ssh_host_dsa_key
Jul 21 10:10:02 nailgun.test.domain.local sshd[690]: Set /proc/self/oom_score_adj from 0 to -1000
Jul 21 10:10:02 nailgun.test.domain.local sshd[690]: error: Bind to port 22 on 10.109.0.2 failed: Cannot assign requested address.
Jul 21 10:10:02 nailgun.test.domain.local sshd[690]: fatal: Cannot bind any address.
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: Child 690 belongs to sshd.service
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: sshd.service: main process exited, code=exited, status=255/n/a
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: sshd.service changed start -> failed
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: Job sshd.service/start finished, result=failed
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: Failed to start OpenSSH server daemon.
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: Unit sshd.service entered failed state.
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: sshd.service failed.
Jul 21 10:10:02 nailgun.test.domain.local systemd[1]: sshd.service: cgroup is empty

[Max Yatsenko (myatsenko)]

Unfortunatelly this issue was not reproduced,

*to reproduce the issue the following configuration was deployed:

MOS8.0 (with last updates)
1Controller
3Computer+Cinder

*After that the following command was run on master node:
fuel-migrate <compute-node-ip>

*Migration proccess was started.

*When migration process was finished , the following message was appeared in master node console:
Source fuelmaster can be obtained by ssh 10.88.0.72

* I was able to log into NEW master node from compute node and run "fule node" command and etc.

As this bug was not reproduced I set status of the bug as Invalid.
Please be free reopen the issue if this problem appears again.