adding host to oVirt issue
by Hetz Ben Hamo
Hi,
I have been checking oVirt latest (4.3.4) with adding various bare metal
hosts with various CPU's: I'm asking on this mailing list if I should open
a ticket and if so what ticket (bug? RFE?).
- AMD Ryzen 5
- Xeon SP (first gen)
- Xeon SP (second gen)
- Upcoming unannounced Xeon CPU.
The main oVirt machine (and the HE) is running on Xeon E3-1225 V3 CPU.
The big problem is that when you add for example any AMD processor
(whatever it's a low end Ryzen 5 or a high end 32 cores EPYC) to such a
setup, it will fail because it tries to add the AMD machine to the default
cluster which is defined as Intel based VT-X. If I'll try to add any non
Xeon V3 CPU, it will succeed, but any live migration of VM's inside the
cluster will fail since any Xeon chip version has various VT flags.
So, my suggestion is that when adding nodes/hosts to oVirt, they will be
added to a new "non clustered" place without any features allowed (live
migration, for example) and after the admin will add the machines, he/she
can assign nodes to a defined cluster (prepared or that he/she can create)
Another problem with cluster defining is that the CPU type is too
complicated for any admin who doesn't know the nuts and bolts of each CPU.
For example - the Haswell CPU type got *6* CPU types. Do you really expect
an admin to know which one to select? IMHO, a good solution would be to add
some "detect" button to hosts, and the result from clicking this button
should be a window with options like:
- Add this host to an existing cluster (which has the same CPU type)
- Auto create a new cluster with the CPU type detected from the
node/host CPU
I'll be happy to hear your thoughts about this issue.
Thanks,
Hetz
2 years
Error: Adding new Host to ovirt-engine
by Ahmad Khiet
Hi,
Can't add new host to ovirt engine, because the following error:
2019-06-12 12:23:09,664 p=4134 u=engine | TASK [ovirt-host-deploy-facts :
Set facts] *************************************
2019-06-12 12:23:09,684 p=4134 u=engine | ok: [10.35.1.17] => {
"ansible_facts": {
"ansible_python_interpreter": "/usr/bin/python2",
"host_deploy_vdsm_version": "4.40.0"
},
"changed": false
}
2019-06-12 12:23:09,697 p=4134 u=engine | TASK [ovirt-provider-ovn-driver
: Install ovs] *********************************
2019-06-12 12:23:09,726 p=4134 u=engine | fatal: [10.35.1.17]: FAILED! =>
{}
MSG:
The conditional check 'cluster_switch == "ovs" or (ovn_central is defined
and ovn_central | ipaddr and ovn_engine_cluster_version is
version_compare('4.2', '>='))' failed. The error was: The ipaddr filter
requires python's netaddr be installed on the ansible controller
The error appears to be in
'/home/engine/apps/engine/share/ovirt-engine/playbooks/roles/ovirt-provider-ovn-driver/tasks/configure.yml':
line 3, column 5, but may
be elsewhere in the file depending on the exact syntax problem.
The offending line appears to be:
- block:
- name: Install ovs
^ here
2019-06-12 12:23:09,728 p=4134 u=engine | PLAY RECAP
*********************************************************************
2019-06-12 12:23:09,728 p=4134 u=engine | 10.35.1.17 :
ok=3 changed=0 unreachable=0 failed=1 skipped=0 rescued=0
ignored=0
whats missing!?
Thanks
--
Ahmad Khiet
Red Hat <https://www.redhat.com/>
akhiet(a)redhat.com
M: +972-54-6225629
<https://red.ht/sig>
2 years
[ OST Failure Report ] [ oVirt Master (vdsm) ] [ 07-06-2019 ] [ 003_00_metrics_bootstrap.metrics_and_log_collector ]
by Dafna Ron
Hi,
We have a failure in vdsm project on master.
The issue is change:
https://gerrit.ovirt.org/#/c/100576/ - Remove SOS VDSM plugin
which is failing on metrics as metrics is calling sos-logcollector.
The patch cannot be changed as until centos 7.7 when sos-3.7-3, which
contains vdsm plugin will come out.
so until then, we are left with no sos plugin, which is causing the metrics
test to fail.
Shirly, can you please take a look and see if we can change the test to not
call sos-logcollector?
Please note, that we are expecting 4.3 to fail on same issue very soon.
failed job can be found here:
https://jenkins.ovirt.org/job/ovirt-master_change-queue-tester/14452/
ERROR from test:
lago.ssh: DEBUG: Command 8626fe70 on lago-basic-suite-master-engine errors:
ERROR: Failed to get a sosreport from:
lago-basic-suite-master-host-1; Could not parse sosreport output
ERROR: Failed to get a sosreport from: lago-basic-suite-master-host-0;
Could not parse sosreport output
lago.utils: DEBUG: Error while running thread Thread-3
Traceback (most recent call last):
File "/usr/lib/python2.7/site-packages/lago/utils.py", line 58, in
_ret_via_queue
queue.put({'return': func()})
File "/home/jenkins/workspace/ovirt-master_change-queue-tester/ovirt-system-tests/basic-suite-master/test-scenarios/003_00_metrics_bootstrap.py",
line 97, in run_log_collector
'log collector failed. Exit code is %s' % result.code
File "/usr/lib64/python2.7/unittest/case.py", line 462, in assertTrue
raise self.failureException(msg)
AssertionError: log collector failed. Exit code is 1
--------------------- >> end captured logging << ----------
Thanks,
Dafna
2 years
Host installation error
by Charles Thao
Hi,
I have a 4.3 engine on Fedora 28 laptop and a 4.3 node running on a
virt-manager. I installed ovirt-node using the latest stable ISO image.
When I tried to connect the host to my engine, this error appeared:
*Host node installation failed. Failed to execute Ansible host-deploy role.*
I've also attached an example log file. Has anyone seen anything similar to
this?
Regards,
Charles T
2 years
Issues when running VDSM tests
by Pavel Bar
Hi Milan,
Can you please check the following VDSM tests's failures:
*Environment:*
1) VDSM master branch;
2) Fedora 29 OS
*Execution (from root folder):*
*make check*
*Issues:*
1) An authentication pop-up window jumps when executing:
*Warning no default label for
/tmp/tmpRCRsYm/libvirtd.............................................................SS.........................................S.SS..S................S.S..................................................................SSS..SSSSS..SSSSSSS.................SSS..SSS...S..............SSSS..SS.....S.........SSSSS.S.S.........................SSSSS....SSSSSS......................................S..SSSSSSSSS.SS.........................................................................................................................................................................----------------------------------------------------------------------*
2) 2 error outputs received, the 2nd one multiple times:
*ERROR: InvocationError for command
'/home/pbar/Dev/git/vdsm/pbar/master/tests/profile network-py27 ./py-watch
600 pytest --durations=5 --cov=vdsm.network
--cov-report=html:htmlcov-network-py27 --cov-fail-under=42
network/integration network/unit' (exited with code 1)*
*E HookError: Hook Error: ('Traceback (most recent call
last):\n File "/usr/libexec/vdsm/hooks/before_nic_hotplug/50_vmfex", line
35, in <module>\n from vdsm import libvirtconnection\nImportError:
cannot import name libvirtconnection\n',)../lib/vdsm/common/hooks.py:124:
HookError*
Thank you in advance!
Pavel
2 years
Building python 2 and python 3 packages for Fedora
by Nir Soffer
In several projects (e.g. vdsm, imageio, sanlock), we have the issue of
building python 3 packages
for Fedora.
The current build process create packages with the same name for both
python 2 and python 3.
When the packages are published to oVirt repository, the python 3 packages
overwrite the python 2
packages.
We can rename the packages properly (e.g. python2-xxx, python3-xxx) but
this requires lot of work
and typically breaks later in the code publishing the packages.
There is also the difficulty of building both python 2 and python 3
packages from same spec in the same
build. This should be possible but not easy.
Since python 2 is about to die soon, should we simplify by building python
2 *or* python 3, depending
on version?
Fedora 29: python 2.7
Fedora 30: python 3.7
CentOS 7: python 2.7
CentOS 8: python 3.6
This make it possible to test and develop on python 2.7 until vdsm is fully
functional on python 3,
and it save resources in the CI.
Thoughts?
Nir
2 years
Failed to add fedora-29 host to oVirt
by Eyal Shenitzky
Hi,
I am trying to add fedora-29 host to me oVirt setup and failed due to the
following error (from the engine log):
2019-06-11 11:24:59,123+03 ERROR
[org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector]
(VdsDeploy) [73c7d83d] EVENT_ID: VDS_INSTALL_IN_PROGRESS_ERROR(511), An
error has occurred during installation of Host fedora_29: DNF Problem:
conflicting requests - package
ovirt-host-4.4.0-0.0.master.20190610132028.git410e453.fc29.x86_64 requires
ovirt-hosted-engine-setup, but none of the providers can be installed -
package ovirt-host-4.4.0-0.0.master.20190610131544.git410e453.fc29.s390x
does not have a compatible architecture - nothing provides
ovirt-host-dependencies = 4.4.0-0.0.master.20190610131544.git410e453.fc29
needed by ovirt-host-4.4.0-0.0.master.20190610131544.git410e453.fc29.s390x
- nothing provides ovirt-hosted-engine-ha >= 2.4 needed by
ovirt-hosted-engine-setup-2.4.0-0.0.master.20190606085519.gitf8402ad.fc29.noarch
- nothing provides vdsm-python >= 4.30 needed by
ovirt-hosted-engine-setup-2.4.0-0.0.master.20190606085519.gitf8402ad.fc29.noarch
- nothing provides ovirt-hosted-engine-ha >= 2.4 needed by
ovirt-hosted-engine-setup-2.4.0-0.0.master.20190610121403.gitb2ee866.fc29.noarch
- nothing provides glusterfs-cli >= 5.6 needed by
ovirt-hosted-engine-setup-2.4.0-0.0.master.20190610121403.gitb2ee866.fc29.noarch
- nothing provides vdsm-python >= 4.40.0 needed by
ovirt-hosted-engine-setup-2.4.0-0.0.master.20190610121403.gitb2ee866.fc29.noarch.
2019-06-11 11:24:59,133+03 ERROR
[org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector]
(VdsDeploy) [73c7d83d] EVENT_ID: VDS_INSTALL_IN_PROGRESS_ERROR(511), An
error has occurred during installation of Host fedora_29: Failed to execute
stage 'Package installation': Problem: conflicting requests - package
ovirt-host-4.4.0-0.0.master.20190610132028.git410e453.fc29.x86_64 requires
ovirt-hosted-engine-setup, but none of the providers can be installed -
package ovirt-host-4.4.0-0.0.master.20190610131544.git410e453.fc29.s390x
does not have a compatible architecture - nothing provides
ovirt-host-dependencies = 4.4.0-0.0.master.20190610131544.git410e453.fc29
needed by ovirt-host-4.4.0-0.0.master.20190610131544.git410e453.fc29.s390x
- nothing provides ovirt-hosted-engine-ha >= 2.4 needed by
ovirt-hosted-engine-setup-2.4.0-0.0.master.20190606085519.gitf8402ad.fc29.noarch
- nothing provides vdsm-python >= 4.30 needed by
ovirt-hosted-engine-setup-2.4.0-0.0.master.20190606085519.gitf8402ad.fc29.noarch
- nothing provides ovirt-hosted-engine-ha >= 2.4 needed by
ovirt-hosted-engine-setup-2.4.0-0.0.master.20190610121403.gitb2ee866.fc29.noarch
- nothing provides glusterfs-cli >= 5.6 needed by
ovirt-hosted-engine-setup-2.4.0-0.0.master.20190610121403.gitb2ee866.fc29.noarch
- nothing provides vdsm-python >= 4.40.0 needed by
ovirt-hosted-engine-setup-2.4.0-0.0.master.20190610121403.gitb2ee866.fc29.noarch.
When I try to install VDSM manually before adding the host the setup it
installed the following version: vdsm-4.18.999-447.git0bb7717.fc28.x86_64
Any suggestions?
--
Regards,
Eyal Shenitzky
2 years
GetDeviceList and direct LUNs
by Fedor Gavrilov
Hi,
I want to implement a feature that would allow one to manually refresh sizes for direct LUNs and I am not sure if existing code in the engine codebase is gonna help with that.
RefreshLunsSizeCommand seems to be doing what is needed, but it operates on a premise that LUNs to refresh belong to a storage domain which makes most of its code irrelevant for purposes of direct LUNs.
It seems that workhorse is GetDeviceList command, but from the code I can't quite understand if it's supposed to be used with external iSCSI targets. Especially 'VDS' part here confuses me since I am not sure what VDS we're talking about in the first place when using direct LUN.
If GetDeviceList is of no use, is there anything else VDSM provides?
Thanks,
Fedor Gavrilov
2 years
