time keeping
by Yaniv Kaul
I'm concerned with two (different) time keeping issues:
1. Timezone - we need to ensure all VMs have the same timezone (At least by
default).
Best option would be to edit the deploy scripts and set the timezone to the
host timezone?
should be fairly easy to add @TIMEZONE@ variable to the deploy script in a
similar manner to what we do today in the domain XML.
It'll be in a command like 'ln -s /usr/share/zoneinfo/@TIMEZONE@
/etc/localtime
2. Time drifting - I have a feeling there's a time drift when we do all the
freeze-thaw and snap/revert... Perhaps we should run 'ntpdate' after such
actions?
9 years
[11-03-2016] Weekly lago release v0.12.0
by David Caro
Hi everyone!
This week we introduced one big change in how lago handles prefixes, we
introduced the 'workdir' concept, that is, a directory that will store one or
more prefixes, let me show with an example.
Imagine that you have an init.yaml that creates an environment with 1 vm, you
can run init like this:
lago --prefix-name one_vm init init.yaml
That will create a workdir at $PWD/.lago with one prefix inside named one_vm
and a soft link named 'current' that points to it, so any further command will
run on that one_vm prefix:
lago status # shows the one_vm prefix path
Until there nothing new, but, now you can create more prefixes, if for example,
if you have another init config file named two_vms.yaml, you can create
another prefix inside that same workdir with:
lago --prefix-name two_vms init two_vms.yaml
That will add a new prefix to $PWD/.lago, it will not update the current link
though, so any command will still run by default on the previous prefix, but
you can change the current prefix with:
lago set-current two_vms
That will make any command that does not specify the --prefix-name option to
use that one by default.
NOTE: lago init will create only workdirs from now on, but you can still use
the old prefixes if you have any around, it will autodetect them as one or the
other.
Some other changes:
Now it will show the libvirt status of the domains
Remove the memory balloon device. We don't use it.
Ignore zero sized RPM files when merging repos.
Added ignore-warnings option
To see the full changelog:
http://resources.ovirt.org/repos/lago/stable/0.0/CHANGELOG_0.0.txt
And as usual, you can get the rpms from:
http://resources.ovirt.org/repos/lago/stable/0.0/rpm/
Cheers!
--
David Caro
Red Hat S.L.
Continuous Integration Engineer - EMEA ENG Virtualization R&D
Tel.: +420 532 294 605
Email: dcaro(a)redhat.com
IRC: dcaro|dcaroest@{freenode|oftc|redhat}
Web: www.redhat.com
RHT Global #: 82-62605
9 years
Fwd: [ovirt-devel] [VDSM] Moving to py.test
by Yaniv Kaul
We probably want to move to it in ovirt-system-tests as well, to align with
both VDSM and Lago itself, no?
---------- Forwarded message ----------
From: Nir Soffer <nsoffer(a)redhat.com>
Date: Thu, Mar 10, 2016 at 3:15 PM
Subject: [ovirt-devel] [VDSM] Moving to py.test
To: devel <devel(a)ovirt.org>, Yaniv Bronheim <ybronhei(a)redhat.com>, Dan
Kenigsberg <danken(a)redhat.com>, Piotr Kliczewski <pkliczew(a)redhat.com>,
Francesco Romani <fromani(a)redhat.com>, Adam Litke <alitke(a)redhat.com>
Hi all,
In the last vdsm call we discussed moving to py.test.
Here is how to run the current tests with py.test:
cd tests
PYTHONPATH=../lib:../vdsm py.test -v *_test.py *_tests.py *Tests.py
network
Currently we have some failures and errors:
8 failed, 1662 passed, 101 skipped, 14 pytest-warnings, 10 error
in 86.05 seconds
Some py.test nice tricks:
Finding the 10 slowest tests:
PYTHONPATH=../lib:../vdsm py.test -v --durations=10 vmTests.py
*_test.py *_tests.py *Tests.py network
Checking code coverage:
PYTHONPATH=../lib:../vdsm py.test -v --cov=../vdsm --cov=vdsm
*_test.py *_tests.py *Tests.py network
Once we finish moving vdsm code to lib/vdsm, renaming the tests to
*_test.py, and add
pytest.ini, running the test will be simply:
py.test
Once all the tests pass, we can:
- Drop all the special asserts (assertXXX), py.test uses plain asserts
https://pytest.org/latest/assert.html#asserting-with-the-assert-statement
- Get better error reports
https://pytest.org/latest/example/reportingdemo.html
- Remove unneeded classes, or inherit from any class
https://pytest.org/latest/getting-started.html
- Replace @permutations with more powerfull @pytest.mark.parametrize
https://pytest.org/latest/parametrize.html#parametrize-basics
- Replace slowtest and stresstest with @pytest.mark
https://pytest.org/latest/mark.html
- Replace monkeypatch.py with simpler and more powerfull pytest
monkeypatch funcarg
https://pytest.org/latest/monkeypatch.html#monkeypatching-mocking-modules...
- Replace NamedTemporaryDir with pytest tmpdir funcarg
https://pytest.org/latest/getting-started.html#going-functional-requestin...
- Simplify complex context managers with pytest fixtures
https://pytest.org/latest/example/simple.html
- Drop testrunner.py, most of testlib.py, some of testValidation.py,
run_tests*.sh
- And more
Nir
_______________________________________________
Devel mailing list
Devel(a)ovirt.org
http://lists.ovirt.org/mailman/listinfo/devel
9 years
Bridge Networking
by Eric Helms
I couldn't find anything in the documentation so asking here, is there a
way to setup bridge networking for a set of VMs to get them public IPs?
Thanks,
Eric
9 years
Fwd: [VDSM] check_merged failures
by Yaniv Bronheim
forwarding to lago devels.. shouldn't be any issues with the tests. nothing
changed since last time it succeeded
---------- Forwarded message ----------
From: Nir Soffer <nsoffer(a)redhat.com>
Date: Tue, Mar 8, 2016 at 4:34 PM
Subject: [VDSM] check_merged failures
To: devel <devel(a)ovirt.org>, Yaniv Bronheim <ybronhei(a)redhat.com>
Hi all,
The new check_merged job is failing randomly. I have see this error today:
14:20:16 @ Start specified VMs:
14:20:16 # Start nets:
14:20:16 * Create network vdsm_functional_tests_lago:
14:20:16 libvirt: error : internal error: Network is already in use
by interface 7f0c-bc616ef
14:20:16 * Create network vdsm_functional_tests_lago: ERROR (in 0:00:00)
14:20:16 # Start nets: ERROR (in 0:00:00)
14:20:16 @ Start specified VMs: ERROR (in 0:00:00)
14:20:16 Error occured, aborting
14:20:16 Traceback (most recent call last):
14:20:16 File "/usr/lib/python2.7/site-packages/lago/cmd.py", line
539, in main
14:20:16 cli_plugins[args.verb].do_run(args)
14:20:16 File
"/usr/lib/python2.7/site-packages/lago/plugins/cli.py", line 180, in
do_run
14:20:16 self._do_run(**vars(args))
14:20:16 File "/usr/lib/python2.7/site-packages/lago/cmd.py", line
132, in wrapper
14:20:16 return func(*args,
prefix=lago_prefix.Prefix(prefix_path), **kwargs)
14:20:16 File "/usr/lib/python2.7/site-packages/lago/cmd.py", line
141, in wrapper
14:20:16 return func(*args, prefix=prefix, **kwargs)
14:20:16 File "/usr/lib/python2.7/site-packages/lago/cmd.py", line
191, in do_start
14:20:16 prefix.start(vm_names=vm_names)
14:20:16 File "/usr/lib/python2.7/site-packages/lago/prefix.py",
line 819, in start
14:20:16 self.virt_env.start(vm_names=vm_names)
14:20:16 File "/usr/lib/python2.7/site-packages/lago/virt.py", line
187, in start
14:20:16 net.start()
14:20:16 File "/usr/lib/python2.7/site-packages/lago/virt.py", line
354, in start
14:20:16 self._env.libvirt_con.networkCreateXML(self._libvirt_xml())
14:20:16 File "/usr/lib64/python2.7/site-packages/libvirt.py", line
4093, in networkCreateXML
14:20:16 if ret is None:raise libvirtError('virNetworkCreateXML()
failed', conn=self)
14:20:16 libvirtError: internal error: Network is already in use by
interface 7f0c-bc616ef
Hopefully someone is owning the infrastructure and can look into this.
Nir
--
*Yaniv Bronhaim.*
9 years
Drop reposync command
by David Caro
Hi everyone,
I'd like to start a discussion about dropping (at least temporarily) the
reposync command.
The reposync command is used in lago for two things:
* Downloads locally all the packages defined in the repo conf passed (yum
config file format)
* Creates an internal_repo dir in the prefix and copies over all the packages
that match the distros used in the prefix
The advantage is that once synced, this allows installing any packages on the
vms without having connection to the internet.
The downside is that right now, it downloads everything, needed or not, so it
takes a lot of space and time the first run to cache all the packages.
For that reason on the ovirt system tests we removed the epel/centos base repos
from the reposync (it was ~20G) but as that already makes the tests require an
internet connection, makes the reposync command obsolete (it is caching only
some part of the packages).
So the idea is to drop the reposync command until we have a better alternative
(maybe a small local proxy that caches everything on the first run or something
similar)
--
David Caro
Red Hat S.L.
Continuous Integration Engineer - EMEA ENG Virtualization R&D
Tel.: +420 532 294 605
Email: dcaro(a)redhat.com
IRC: dcaro|dcaroest@{freenode|oftc|redhat}
Web: www.redhat.com
RHT Global #: 82-62605
9 years
Was the check for 'lago' group membership removed?
by Yaniv Kaul
I uninstalled and reinstalled lago, which made the lago group
disappear and reapper - and my user was not part of it.
Then running the tests failed, as I could not write the lock file to
/var/lib/lago/...
Added a bit of code to the exception to see why:
File "/usr/lib/python2.7/site-packages/lockfile/linklockfile.py",
line 20, in acquire
raise LockFailed("Failed to create %s.\nCode: %d\nReason: %s\n" %
(self.unique_name, e.errno, e.strerror))
LockFailed: Failed to create
/var/lib/lago/store/ykaul-mini.tlv.redhat.com-fa517700.1377-7176078113883238881.
Code: 13
Reason: Permission denied
And indeed, it was because I was not part of the lago group. Adding,
logging out and in again solved it - but shouldn't there be such
warning ? (I remember there was in the past).
TIA,
Y.
9 years, 1 month
'make check' fails due to yaml
by Yaniv Kaul
After today's (?) changes:
tests/unit/test_config.py::test_nonexistent_throws PASSED
=====================================================================================================
ERRORS ======================================================================================================
___________________________________________________________________________________
ERROR collecting tests/unit/test_dirlock.py
___________________________________________________________________________________
tests/unit/test_dirlock.py:27: in <module>
from lago import dirlock
lago/dirlock.py:25: in <module>
import utils
lago/utils.py:36: in <module>
import yaml
E ImportError: No module named yaml
_________________________________________________________________________________
ERROR collecting tests/unit/lago/test_prefix.py
_________________________________________________________________________________
tests/unit/lago/test_prefix.py:6: in <module>
from lago import prefix
lago/prefix.py:36: in <module>
import subnet_lease
lago/subnet_lease.py:45: in <module>
import utils
lago/utils.py:36: in <module>
import yaml
E ImportError: No module named yaml
_________________________________________________________________________________
ERROR collecting tests/unit/lago/test_utils.py
__________________________________________________________________________________
tests/unit/lago/test_utils.py:2: in <module>
import yaml
E ImportError: No module named yaml
========================================================================================
7 passed, 3 error in 0.08 seconds
========================================================================================
Makefile:52: recipe for target 'check-local' failed
make: *** [check-local] Error 1
I'm on:
commit 5178f92f435791a778d4fbd801baab532819e7b0
Merge: 5deacae 79bb6ca
Author: Lago CI bot <dcaroest+cibot(a)redhat.com>
Date: Tue Mar 1 16:38:23 2016 +0100
Merge pull request #151 from lago-project/yaml_init_definition
Added yaml support for the init config file
9 years, 1 month
