Adding the Lago devel mailing list.
The download is the reposync phase - which seems to be OK, but then the
connection means that for some reason Lago is not serving those RPMs (8585
is the port it should be listening to).
Can you share some logs?
On Wed, Sep 21, 2016 at 4:25 PM, Nicolas Ecarnot <nicolas(a)ecarnot.net>
> Hello Yaniv,
> Following your recommendation, I install a new bare metal host, installed
> F24, then followed the lago.readthedoc web site.
> When running ./run_suite.sh basic_suite_3.6 , things seems to go right
> (downloading...) for a while, then begin some errors about accessing
> May you tell me where would be the best place to ask some help about Lago
> ? (IRC, mailing list, e-mail... ?)
> Best regards,
> Nicolas ECARNOT
Le 26/09/2016 à 12:04, Yedidyah Bar David a écrit :
> On Mon, Sep 26, 2016 at 1:02 PM, Nicolas Ecarnot <nicolas(a)ecarnot.net> wrote:
>> Le 26/09/2016 à 12:00, Yedidyah Bar David a écrit :
>>> On Mon, Sep 26, 2016 at 12:36 PM, Nicolas Ecarnot <nicolas(a)ecarnot.net>
>>>> Le 26/09/2016 à 10:55, Yedidyah Bar David a écrit :
>>>>> So we need engine logs to know more.
>>>>> If none were generated, it might have failed too early.
>>>>> Please check system logs - /var/log/messages, journalctl, etc. Thanks.
>>>> Just to be sure : are you speaking about the log files on the bare-metal
>>>> server, or the virtual hosts (engine)?
>>> I referred to the engine vm. In principle both might be
>>>> If this is the engine, I already provided the only file I got (and I also
>>>> double-checked its /var/log/message, with no interesting info).
>>> Perhaps journalctl output?
>> Engine is centos 6, so no systemd at this time...
> OK. Can you try starting the service manually and see if you get anything
> in any log? You can try getting more debug info using e.g.:
I read this BZ, and I also tried to launch a run using basic_suite_4.0
The results were quite different, as I see no more memory issues.
Actually, it looks like all the 3 VMs are deployed OK (engine, host0 and
Then the tests about the engine are OK.
Then I only have an error in post-001_initialize_engine.py
But I can not find any detail :
@ Collect artifacts:
# [Thread-1] lago_basic_suite_4_0_host0:
# [Thread-2] lago_basic_suite_4_0_host1:
# [Thread-3] lago_basic_suite_4_0_engine:
No handlers could be found for logger "paramiko.transport"
# [Thread-3] lago_basic_suite_4_0_engine: Success (in 0:01:07)
# [Thread-1] lago_basic_suite_4_0_host0: ERROR (in 0:02:18)
# [Thread-2] lago_basic_suite_4_0_host1: ERROR (in 0:02:18)
The script is not stopping there, and keeps going until 002_bootstrap.py :
# add_dc: Success (in 0:00:03)
# add_dc_quota: Success (in 0:00:02)
# add_cluster: Success (in 0:00:01)
* Collect artifacts:
* Collect artifacts: ERROR (in 0:01:55)
# add_hosts: ERROR (in 0:17:08)
# Results located at
@ Run test: 002_bootstrap.py: ERROR (in 0:17:16)
Error occured, aborting
Traceback (most recent call last):
File "/usr/lib/python2.7/site-packages/ovirtlago/cmd.py", line 258,
File "/usr/lib/python2.7/site-packages/lago/plugins/cli.py", line
180, in do_run
File "/usr/lib/python2.7/site-packages/lago/utils.py", line 488, in
return func(*args, **kwargs)
File "/usr/lib/python2.7/site-packages/lago/utils.py", line 499, in
return func(*args, prefix=prefix, **kwargs)
File "/usr/lib/python2.7/site-packages/ovirtlago/cmd.py", line 102,
raise RuntimeError('Some tests failed')
RuntimeError: Some tests failed
I'm still having a look at the log files to see what is the issue.
Apart that, by connecting into the engine VM, I saw that the engine
process was running, so I tried to access the web GUI, by running an SSH
connection to the bare-metal host :
ssh -L 8443:192.168.200.4:443 root(a)serv-hv-dev01.sdis.isere.fr
(BTW, the readthedoc is showing 200.2 but this is irrelevant for
basic_test_4.0, where the engine ip is 200.4)
Accessing https://localhost:8443/ is working, but when trying to access
the login screen, I'm left with :
"The client is not authorized to request an authorization. It's required
to access the system using FQDN."
Dear Redhat employees and al., I sincerely respect your great patience
because I'm witnessing the huge complexity of this project, as I'm
trying to find my way out this mess^W wonderful project ;)
Anyway, I'm feeling I made one step further, so this is encouraging (I
have modest expectations and a few is making my day).
So about this last issue, please don't answer me that every Lago user is
using its own laptop with a graphical layer and running a direct browser
access to the web GUI?
I've noticed a strange issue today. After running successfully lago
init && start, then running lago stop && destroy, the qemu-kvm VM's
processes are still running even though the destroy command completed
Anyone else encountered this?
I've created a new RTD page for oVirt-system-tests  as its an
independent project and shouldn't be part of the official Lago
documentation, especially not as 1st how-to example.
Please feel free to contribute to it and update it with new info,
especially on adding new tests or devel use.
I've also sent a new PR  to remove most of the OST and link to the new
OST documentation page, and replacing the main example with the Lago
Going forward, we'll probably create a new lago-examples.git and reference
to it for all examples documentation.
You'll also notice I've removed some of the documentation around advanced
debugging and explanations on internal Lago structure which shouldn't be on
how how to page, but in the Lago Tutorial which is also in progress at .
I believe this will help new users get Lago up and running much faster and
not handle complex settings like nested virt if they are not interested in
installing oVirt at first.
EMEA ENG Virtualization R&D
Red Hat Israel
irc: eedri (on #tlv #rhev-dev #rhev-integ)
Hi lake swimmers,
In this pr I am adding an ansible playbook (which will grow to a role)
to install lago and system-tests. Brought up by maintainers and duck-rh,
its more appropriate to create an ansible focused repo which will also play
nicely with galaxy and will contain potentially multi roles for all sorts
of automation and orchestration we can come up with for lago and lago
Please create `ansible-role-lago` and I'll move the below pr to it.
As discussed in the PR, we might need to think of restructuring our
I suggest the following structure:
- lago-project/lago-examples.git - Examples of how to use lago to deploy 'X'
This repository will hold all examples of deploying Lago, each example
will have its own directory. Currently we have:
lago-project/lago-examples/jenkins - the current lago-demo
repository(scripts to deploy Jenkins using Lago)
lago-project/lago-examples/ovirt-system-tests - installing lago +
The same as 'lago-examples.git', only it will hold ansible playbooks
that use lago and ansible for deployment(such as the PR)
Holding a generic ansible playbook to install lago(once we have one :)
Though I'll note that I have no concrete idea of what are the ansible
As a side-effect of my proposed PR, 3 backward-incompatible
changes will be added:
variables names in lago.conf
1. 'default_ssh_user', changed to: 'ssh_user', value: no change.
2. 'default_ssh_password, changed to: 'ssh_password', value: no change.
3. 'template_default_repo', changed to: 'template_repo_name'(same as
the existing CLI), value: was never set.
The above 3 were hard-coded in config.py's defaults dictionary and
never exposed externally. They had no documents, but, if a user knew
about them he could have added them manually to lago.conf
The reason for changing is that the PR unifies the configuration files
and CLI parameters. So keeping the 'default' prefix is misleading and
will lead to awkward results, such as:
lago --default-ssh-user root --default-ssh-password ....
lago --ssh-user root --ssh-password ...
'template_*' parameters which are used by lago init command, will be
move to [init] section in lago.conf.
1. Drop /etc/lago.d/lago.conf in favour of /etc/lago/lago.conf
2. The installed file from the RPM will be all commented out(i.e.
installed just to use as a reference).
3. /etc/lago/lago.conf will be the least important in the
hierarchy(you can read more details about the hierarchy in the PR)
I'll note that Lago will be fully functional without any configuration
file at all.
These changes will possibly affect only already-customized
lago.conf files. But, changing to the new configuration format will be
easy, using the 'generate-config' command:
mkdir -p ~/.config/lago && lago generate-config > ~/.config/lago/lago.conf
The changes are needed as the the PR mainly deals with standardizing
the configuration loading module, so it is a good time for clean-up,
and dropping misleading variable names(which were probably good at the
time written). Second, I'm not sure how widely these parameters were
used, as some of them didn't have any docs. For instance I couldn't
find any evidence of usage for 'template_default_repo'(the CLI
parameter is the same as it was, 'template_repo_name').
Some possible alternatives:
1. Issue a warning about variables name change, and use both, then in
2-3 versions from now, completely remove them.
2. lago.conf location: Not a major issue, as far as I know, '.d'
suffix was used historically for distinguishing a 'directory' from a
single configuration file, so having '/etc/lago/' is better(and
keeping lago.d if someday we will need to load other user-defined
configurations). But this can be easily changed.
If there is no request for the alternatives or objection, I'll try to
get it merged for the next version and add a short summary of this
message to the release email.
Lago 0.25 is out, introducing few bug fixes, and 2 new features:
- /var/lib/lago/subnets is now configurable, this is useful if you
want to install Lago in virtualenv(thanks @pilou-)
- virtio-scsi support for none-root disks. More details in the PR
Here is the full change log:
* Wed Sep 21 2016 Lago CI bot <dcaroest+cibot(a)redhat.com> - 0.25.0
fdd63ae4: Merge pull request #298 from pilou-/make_lease_dir_configurable
5fb60899: Make lease_dir configurable
* Tue Sep 20 2016 Lago CI bot <dcaroest+cibot(a)redhat.com> - 0.24.5
8845853f: Merge pull request #323 from mykaul/get_cpu_caps
350db807: Merge branch 'master' into get_cpu_caps
83a7f68a: Add host CPU model detection
* Tue Sep 20 2016 Lago CI bot <dcaroest+cibot(a)redhat.com> - 0.24.4
1d7722a6: Merge pull request #322 from rafaelmartins/master
9f423604: utils: fixed typo
* Tue Sep 20 2016 Lago CI bot <dcaroest+cibot(a)redhat.com> - 0.24.3
3dd77113: Merge pull request #318 from lago-project/add_boot_menu
26e9dd71: Adding boot menu
* Sun Sep 11 2016 Lago CI bot <dcaroest+cibot(a)redhat.com> - 0.24.2
061207c6: Merge pull request #314 from mykaul/add_serial_to_disks
c0baab76: Add serial number to disks.
* Wed Sep 07 2016 Lago CI bot <dcaroest+cibot(a)redhat.com> - 0.24.1
27856ea9: Merge pull request #306 from mykaul/virtio-scsi
af1c52d6: virtio-scsi support in Lago
6334cc2b: Merge branch 'master' into virtio-scsi
Repo - http://resources.ovirt.org/repos/lago/stable/0.0/rpm/
Docs - http://lago.readthedocs.io/en/0.25/Env_Setup.html
Is it possible to change the generated hostnames in Lago - currently it is
of the form lago_basic_suite_hc_host1 and using this hostname fails while
peer probing gluster.
According to gluster dev:
"The valid_host_name () function what we have at gluster CLI is as per RFC
1912 and an underscore in the hostname is *not* accepted."
If not, other suggestions?
Not sure if this is the right forum for this newbie question - if not,
please point me to the right one.
My issue is the repos available in the guest VMs started by lago.
I have executed from the working dir
#lagocli ovirt reposetup
When I launch into shell of guest VM :
[root@lago_basic_suite_hc_host0 ~]# yum repolist
Loaded plugins: fastestmirror
http://192.168.201.1:8585/el7/repodata/repomd.xml: [Errno 14] curl#7 -
"Failed connect to 192.168.201.1:8585; Connection refused"
Trying other mirror.
I must be missing some step, could you help me figure out what?