[ovirt-devel] Re: Problem on host deployment from engine

On Mon, Apr 15, 2019 at 3:14 PM Nir Soffer <nsoffer(a)redhat.com&gt; wrote: > > > > On Mon, Apr 15, 2019, 14:59 Amit Bawer <abawer(a)redhat.com&gt; wrote: >> >> Hello Didi & Sandro, >> >> I have encountered following issue when attempting to deploy a host from the engine management. >> >> Engine: Fedora 28 >> Host: CentOS 7.6.1810 >> >> engine.log error: >> >> 2019-04-14 14:08:35,578+03 INFO [org.ovirt.engine.core.uutils.ssh.SSHDialog] (EE-ManagedThreadFactory-engine-Thread-2) [628278a7] SSH execute ' root(a)10.35.0.229&#39; 'umask 0077; MYTMP="$(TMPDIR="${OVIRT_TMPDIR}" mktemp -d -t ovirt-XXXXXXXXXX)"; trap "chmod -R u+rwX \"${MYTMP}\" > /dev/null 2>&1; rm -fr \"${MYTMP}\" > /dev/null 2>&1" 0; tar --warning=no-timestamp -C "${MYTMP}" -x && "${MYTMP}"/ovirt-host-deploy DIALOG/dialect=str:machine DIALOG/customization=bool:True' >> 2019-04-14 14:08:35,670+03 ERROR [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] (VdsDeploy) [628278a7] EVENT_ID: VDS_INSTALL_IN_PROGRESS_ERROR(511), An error has occurred during installation of Host host1: Python is required but missing. >> 2019-04-14 14:08:35,688+03 ERROR [org.ovirt.engine.core.uutils.ssh.SSHDialog] (EE-ManagedThreadFactory-engine-Thread-2) [628278a7] SSH error running command root@10.35.0.229:'umask 0077; MYTMP="$(TMPDIR="${OVIRT_TMPDIR}" mktemp -d -t ovirt-XXXXXXXXXX)"; trap "chmod -R u+rwX \"${MYTMP}\" > /dev/null 2>&1; rm -fr \"${MYTMP}\" > /dev/null 2>&1" 0; tar --warning=no-timestamp -C "${MYTMP}" -x && "${MYTMP}"/ovirt-host-deploy DIALOG/dialect=str:machine DIALOG/customization=bool:True': IOException: Command returned failure code 1 during SSH session &#39;root(a)10.35.0.229&#39; >> >> >> I managed to resolve it manually by installing following rpms over the host machine before reattempting to deploy again the host from the engine management: >> >> python2-otopi >> python2-ovirt-host-deploy On the _host_ (the hypervisor, the machine you want to add as a host to the engine), or on the _engine_ machine? If former, please try latter and report. Thanks. You should not need to install anything on the host machine. If you do, that's a bug.

For adding a python2 host, you need python2 otopi/host-deploy packages on the engine machine. That's expected. If that's your only issue, and we want to "solve" it, we have two options. I personally do not have a strong preference: 1. Merely document this somewhere (in engine's README or whatever) 2. Make the engine require both python2-otopi and python3-otopi Latter option will very cleanly and easily solve your current problem, but will not be possible once we support the engine on an OS that has only python3 - perhaps fedora 31 or so, see e.g.: https://fedoraproject.org/wiki/Changes/Mass_Python_2_Package_Removal

Best regards, -- Didi