This could be related to jdk bug.

I don't really understand the problem, nice @Nir Soffer reports that Engine running on el7 is able to ssh into fc28 host.

If you can hack ssh config on Engine side, it would be great.

If hacking on the host is needed, it is less great, but it is still better to hack it, file a bug, and get on with OST.

The only thing that came to my mind is to enable the same SSH
ciphers/algorithms on FC28 which are enabled on EL7 now

On Wed, 27 Jun 2018, 11:24 Gal Ben Haim (oVirt JIRA), <

@Martin Perina Is there any temporary workaround for this issue?
I really want to go forward and discover other issues that are related to FC28.

For reference, bugzilla is https://bugzilla.redhat.com/show_bug.cgi?id=1591801

We already have a bug for that issue, we need to upgrade apache ssh java
library from 0.12 to 1.y, which is a huge API change. Btw we have already
tried that upgrade and we failed, so it is not something we will be able to
fix in a few days, but we will put a priority on that in upcoming weeks.

Martin

On Tue, 26 Jun 2018, 16:40 Gal Ben Haim (oVirt JIRA), <

Sorry for the distraction. vdsm was not even installed on host-1:

2018-06-26 07:12:41,218-04 DEBUG [org.ovirt.engine.core.uutils.ssh.SSHClient] (default task-17) [ec8e0859-5d6e-48c9-90b9-e6ffb35ebc8a] Connecting 'root@lago-basic-fc-suite-master-host-1'
2018-06-26 07:12:41,455-04 DEBUG [org.ovirt.engine.core.uutils.ssh.SSHClient] (default task-17) [ec8e0859-5d6e-48c9-90b9-e6ffb35ebc8a] Connected: 'root@lago-basic-fc-suite-master-host-1'
2018-06-26 07:12:41,456-04 ERROR [org.ovirt.engine.core.uutils.ssh.OpenSSHUtils] (default task-17) [ec8e0859-5d6e-48c9-90b9-e6ffb35ebc8a] The key algorithm 'EC' is not supported, will return null.
2018-06-26 07:12:41,462-04 ERROR [org.ovirt.engine.core.bll.hostdeploy.AddVdsCommand] (default task-17) [ec8e0859-5d6e-48c9-90b9-e6ffb35ebc8a] Failed to establish session with host 'lago-basic-fc-suite-master-host-1': null
2018-06-26 07:12:41,462-04 WARN [org.ovirt.engine.core.bll.hostdeploy.AddVdsCommand] (default task-17) [ec8e0859-5d6e-48c9-90b9-e6ffb35ebc8a] Validation of action 'AddVds' failed for user admin@internal-authz. Reasons: VAR_ACTIONADD,VARTYPE_HOST,$server lago-basic-fc-suite-master-host-1,VDS_CANNOT_CONNECT_TO_SERVER

on host-0 it was installed by means of
2018-06-26T11:07:14Z DDEBUG Command: dnf -y install ovirt-host

on neither host do I see any attempt to start vdsm, most probably due to the failed ssh.

Open

@danken

https://jenkins.ovirt.org/job/ovirt-system-tests_standard-check-patch/1349/artifact/exported-artifacts/check-patch.basic_fc_suite_master.el7.x86_64/test_logs/basic-fc-suite-master/post-002_bootstrap.py/lago-basic-fc-suite-master-host-0/_var_log/vdsm/ shows that vdsm.log is empty, meaning that vdsmd has not started.

Debugging further is difficult, since /var/log/mesasges is missing and the journal is an ugly binary blob
https://jenkins.ovirt.org/job/ovirt-system-tests_standard-check-patch/1349/artifact/exported-artifacts/check-patch.basic_fc_suite_master.el7.x86_64/test_logs/basic-fc-suite-master/post-002_bootstrap.py/lago-basic-fc-suite-master-host-0/_var_log/journal/

Can you look inside it for a reason why vdsm is down?