In 4.3 we are dropping fluentd in favor of pcp and rsyslog (bug #1627753). Removing fluentd in 4.3 hosts means that 4.2 engine need to be able to add a 4.3 host to a 4.2 cluster in 4.2 compatibility mode without failing even if fluentd is missing. If this can't be achieved we'll need to ship both fluentd and pcp + rsyslog on RHV-H 4.3 so adding pcp and rsyslog side by side with existing fluentd. Tentatively targeting to 4.2.7 at least for initial investigations.
Adding 4.3 host without fluentd to 4.2 engine (with metrics setup) failed on ansible role 2019-01-16 10:25:16,373+01 ERROR [org.ovirt.engine.core.bll.hostdeploy.InstallVdsInternalCommand] (EE-ManagedThreadFactory-engine-Thread-124381) [4e10f9f6] Host installation failed for host '54dd97c5-41cc-47ca-8268-824ea57d2ba4', 'vhost09': Failed to execute Ansible host-deploy role. Please check logs for more details: /var/log/ovirt-engine/host-deploy/ovirt-host-deploy-ansible-20190116102342-10.37.140.251-4e10f9f6.log in ansible log: 2019-01-16 10:25:09,513 p=24221 u=ovirt | TASK [oVirt.ovirt-fluentd/ovirt-fluentd-package : Install fluentd package] ***** 2019-01-16 10:25:16,149 p=24221 u=ovirt | fatal: [10.37.140.251]: FAILED! => { "changed": false, "rc": 126, "results": [ "No package matching 'fluentd' found available, installed or updated" ] } Tested in ovirt-engine-4.2.8.2-0.1.el7ev.noarch with ovirt-engine-metrics-1.1.8-1.el7ev.noarch, ansible-2.7.5-1.el7ae.noarch
(In reply to Lucie Leistnerova from comment #1) > Adding 4.3 host without fluentd to 4.2 engine (with metrics setup) failed on > ansible role Was the cluster set to 4.2 compatibility?
What if engine is 4.3, compatibility set to 4.2, adding a 4.3 RHV-H?
Yes, the compatibility on cluster was 4.2. I'll test the scenario for upgrade to 4.3 and when there will be a problem I'll create new BZ.
I tested the compatibility of the scenario for upgrade to 4.3 and everything is working. Steps: 1) Have a working metrics store and 4.2 engine with 4.2 hosts 2) Upgrade engine to 4.3 and test if metrics are still working 3) Upgrade hosts to 4.3 and test if metrics are still working Result: Metrics store is still working and collecting data. Upgrade to 4.3 tested with: ovirt-engine-4.3.0-0.8.rc2.el7.noarch, ovirt-engine-metrics-1.2.0.1-1.el7ev.noarch, ansible-2.7.5-1.el7ae.noarch
Host 4.3 still can't be added with the same error. I don't see new version of ovirt-engine-metrics in our testing build, only ovirt-engine-metrics-1.1.8-1.el7ev.noarch from rhel-7-server-rhv-manager repo.
The build does not include ovirt-engine-metrics 1.1.8.1?
Now is 1.1.8.1 available and 4.3 host was added successfully. verified in ovirt-engine-metrics-1.1.8.1-1.el7ev.noarch with ovirt-engine-4.2.8.3-0.1.el7ev.noarch
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2019:0358