July 2016 - Devel - oVirt List Archives

gerrit.ovirt.org migration
by Shlomo Ben David 10 Jul '16

10 Jul '16

Hi All, Today we are planning to migrate gerrit.ovirt.org server to a new bigger instance on Amazon to improve performance. - The migration will start at 18:00 IDT - Estimate migration end time 19:00 IDT. - During the migration the server will not be available (you won't be able to send patches or review code) - The server will be down for about an hour, but if we'll be able to restore it before we'll let you know. - An email will be sent at the end of the migration process. Best Regards, Shlomi Ben-David | Software Engineer | Red Hat ISRAEL Phone: +972-54-8008858 IRC: shlomibendavid (on #rhev-integ, #rhev-dev, #rhev-ci) OPEN SOURCE - 1 4 011 && 011 4 1

6 7

oVirt master moving to F24
by Eyal Edri 10 Jul '16

10 Jul '16

FYI, We encourage all maintainers to switch their projects builds for master to F24, as oVirt project builds usually only latest release (unless there are issues that prevents it). ovirt-engine & mom already moved, this usually requires a single patch to jenkins yaml file for your standard CI project, see [1] for e.g on engine and also you'll need to update the master publisher to take the f24 rpms instead, see [2] for e.g feel free to contact infra list if you need help. [1] https://gerrit.ovirt.org/#/c/60232/ [2] https://gerrit.ovirt.org/#/c/60392/ -- Eyal Edri Associate Manager RHEV DevOps EMEA ENG Virtualization R&D Red Hat Israel phone: +972-9-7692018 irc: eedri (on #tlv #rhev-dev #rhev-integ)

1 0

(MASTER) exception on server.log: ERROR: insert or update on table "host_device" violates foreign key constraint "fk_host_device_parent_name"
by Yaniv Kaul 07 Jul '16

07 Jul '16

Caused by: org.postgresql.util.PSQLException: ERROR: insert or update on table "host_device" violates foreign key constraint "fk_host_device_parent_name" Detail: Key (host_id, parent_device_name)=(767142d5-6242-4bb1-bcbd-095595c3b7f1, scsi_host2) is not present in table "host_device". at org.postgresql.core.v3.QueryExecutorImpl.receiveErrorResponse(QueryExecutorImpl.java:2157) at org.postgresql.core.v3.QueryExecutorImpl.processResults(QueryExecutorImpl.java:1886) at org.postgresql.core.v3.QueryExecutorImpl.execute(QueryExecutorImpl.java:255) at org.postgresql.jdbc2.AbstractJdbc2Connection.executeTransactionCommand(AbstractJdbc2Connection.java:793) at org.postgresql.jdbc2.AbstractJdbc2Connection.commit(AbstractJdbc2Connection.java:817) at org.jboss.jca.adapters.jdbc.local.LocalManagedConnection.commit(LocalManagedConnection.java:96) Seen on ovirt-engine-4.1.0-0.0.master.20160701111322.gitf1c9a5c.el7.centos.noarch Y.

3 3

[FYI] Jenkins maintenance today 18:30 TLV
by Nadav Goldin 06 Jul '16

06 Jul '16

Hi, http://jenkins.ovirt.org will be restarted for plugin updates today at 18:30 TLV time, expected downtime is 30 minutes. Patches sent 30 minutes before might not get checked, patches sent during the downtime will get checked when Jenkins is back. If patches you sent did not trigger CI, you can login after the downtime and re-trigger them manually. Thanks, Nadav.

1 1

oVirt Community Newsletter: June 2016
by Brian Proffitt 06 Jul '16

06 Jul '16

Not only was June the occasion of the Red Hat Summit event in San Francisco, where oVirt was well-represented in the event and within Community Central, but it was also the month where the year's biggest release, oVirt 4.0! The latest release of oVirt features: * A New Administration Portal * Improved Live-Migration Performance * Improved Image Features * Container Support * New and Improved oVirt Node Check out oVirt 4.0 today, and in the meantime, here's what happened in June, 2016: ----------------- Software Releases ----------------- oVirt 4.0.0 Final Release is now available http://lists.ovirt.org/pipermail/announce/2016-June/000268.html New oVirt-Live (4.0.0) is available for download http://lists.ovirt.org/pipermail/users/2016-June/040577.html oVirt 3.6.7 Fifth Release Candidate is now available for testing http://lists.ovirt.org/pipermail/announce/2016-June/000265.html ---------------- In the Community ---------------- oVirt 4.0 is released! http://www.ovirt.org/blog/2016/06/ovirt-40-release/ Go Upstream at Summit Community Central http://red.ht/29ipa6z oVirt Meetup Boston - MA http://www.meetup.com/Boston-oVirt-Community/ A New oVirt Dashboard for Disaster Recovery https://bitbucket.org/chocomango/ovirt-dashboard P.L. Ferrari Deploys Red Hat Enterprise Virtualization for a More Efficient IT Infrastructure http://red.ht/1V12049 Red Hat Summit Community Track Day One http://red.ht/29mAH5m oVirt 4.0 mit neuem Dashboard (Deutsch) http://www.pro-linux.de/news/1/23702/ovirt-40-mit-neuem-dashboard.html ------------------------------------ Deep Dives and Technical Discussions ------------------------------------ Supercharge Your Network Throughput via Single Root I/O Virtualization (SR-IOV) http://red.ht/1RX3Qkx oVirt Release 4.0: New opportunities [Russian] [Video] https://youtu.be/I4hQAH08Dlg How to Install oVirt (Open Virtualisation) on Centos 7 [Indonesian] [Video] https://youtu.be/eO0tzmQ9LCk Fedora 24 Server & oVirt 4.0 [Video Playlist] http://bit.ly/29iqDtk Creating a KVM VM through oVirt 4.0 with Intel Skull Canyon [Video] https://youtu.be/AjmstE30sM0 Unedited Install of Windows 2012 R2 on KVM oVirt with Nuctastic Intel Skull Canyon [Video] https://youtu.be/DnhutZdLYT4 3 Minute Provisioning VM from KVM oVirt Template [Video] https://youtu.be/uH7QraPPFnw Restart of KVM Ovirt Windows 2016 VM in less than 2 seconds [Video] https://youtu.be/Ea_QeOv9mD8 -- Brian Proffitt Principal Community Analyst Open Source and Standards @TheTechScribe 574.383.9BKP

1 0

SPM or SDM for a new verb?
by Shmuel Melamud 05 Jul '16

05 Jul '16

Hi! I'm writing code for a new verb (sparsifyInplace) in VDSM and got two different opinions about whether to use SPM or SDM for it: 1) SDM is the new correct approach, need to use it. 2) SDM is on early stage and may be changed significantly, so it is better to use SPM as mature and reliable approach. What's your opinion? Shmuel

2 1

Ovirt Node Next 4.0.0 host networking setup attempts to bridge an Infiniband interface
by Giorgio Biacchi 05 Jul '16

05 Jul '16

Hello list, currently I have an up and running Ovirt 3.6 installation where every hypervisor access the data domain via NFS over infiniband. All is working as expected, with a couple of hooks I can set connected mode on ib0 and an MTU of 65520. Now I'm testing Ovirt Node Next, but the behavior is different. When I try to set up host networking the procedure attempts to bridge the network "ibstorage", defined in Networks tab, with ib0 interface but a non ethernet interface cannot be bridged that way. There's any way I can obtain the desired result? Or somehow tell to the hypervisor to not bridge that connection and to use ib0 directly? Thanks -- gb PGP Key: http://pgp.mit.edu/ Primary key fingerprint: C510 0765 943E EBED A4F2 69D3 16CC DC90 B9CB 0F34

2 1

update - no infra support for the next 2 days
by Eyal Edri 05 Jul '16

05 Jul '16

FYI, Due to various holidays, PTOs and an on going 3.6.8 build there is not infra support for the next two days. If there are urgent issues, please open a ticket to infra-support(a)ovirt.org and we'll do our best to handle it if possible. -- Eyal Edri Associate Manager RHEV DevOps EMEA ENG Virtualization R&D Red Hat Israel phone: +972-9-7692018 irc: eedri (on #tlv #rhev-dev #rhev-integ)

1 0

Integration tests future (and very nice alternative for the DAO fixture file)
by Roman Mohr 04 Jul '16

04 Jul '16

Hi all, In [1] you can find some patches which are meant to improve the test writing experience in ovirt-engine. They provide the following things: A) Domain Object builders which can be used for creating and/or persisting domain objects [2] B) DAO testing without writing fixtures because of the builders C) Integration testing for commands in conjunction with a real database Arquillian, injectable commands and the builders [3] # How to run what? A) In normal unit tests just create a new instance of a builder and use it. This should help us to get rid of all the small createDefaultVm(), createHostWithX() helper methods in our tests. B) In dao tests just inject them and go ahead. The advantage of not using the fixture file is that we can now set up clean scenarios for every test in a setup method. See example 2 below on how easy it is to set up a new cluster. C) Arquillian integration tests need to be marked with "@Category(IntegrationTest.class)" and can inherit from TransactionalTestBase. The @Category annotation makes sure that the integration tests are only run when mvn clean verify -DskipITs=false is invoked. Note that these tests are then executed in the integration test phase of maven. For them we use the maven-failsafe-plugin[5] which will also make sure that the testing database is up to date. See [4] for more details. # Examples 1) Add a running VM to a host, persist everything to the database and load all VMs which are running on the host: VDS host = vdsBuilder.cluster(persistedCluster).persist(); vmBuilder.host(host).up().persist(); List<VM> vms = vmDao.getAllRunningForVds(host.getId()); 2) Add 10 hosts with 1 GB of RAM to a cluster, persist the hosts to the database in a DAO test: public class MyHostDaoTest extends BaseDaoTestCase { @Inject private VdsBuilder vdsBuilder; @Test public void createHosts() { VdsBuilder builder = vdsBuilder.cluster(persistedCluster).physicalMemory(1000); for (int x =0; x < 10; x++){ builder.id(Guid.newGuid()).persist(); } } } 3) Full integration test with arquillian and the database @Category(IntegrationTest.class) public class VmDaoIntegrationTest extends TransactionalTestBase { @Inject VmDao vmDao; private final Guid VM1_GUID = Guid.createGuidFromString("0fe4bc81-5999-4ab6-80f8-7a4a2d4bfacd"); @Deployment public static JavaArchive deploy(){ return createDeployment(); } @Test public void shouldFailOnExistingEntity() { vmBuilder.id(VM1_GUID).cluster(clusterBuilder.reset().persist()).persist(); // This uses assertThat from assertj: assertThat(vmDao.get(VM1_GUID)).isNotNull(); } } 4) Using the builders in a normal unit test without a database: VM vm = new VmBuilder().id(Guid.newGuid()).up().build(); # How to add your own Domain objects? There are just a few simple rules: 1) Your builder should extend org.ovirt.engine.core.builder.AbstractBuilder 2) Make sure that you only access DAOs injected into the builder during #prePersist() and #persist(). This allows to use the #build() method also without injections 3) #prePersist() should set all fields which are necessary to suffice database constraints. The fields should only be set if they are not already set before by the builder. When following this rule we can always persist new objects to the database by simply calling myBuilder.reset().persist(). 4) Mark your builder with @Repository to make them useable for our Spring DAO tests and our Arquillian integration tests. So have a look at the patches at [1] and let me know what you think about them. Best Regards and happy testing, Roman [1] https://gerrit.ovirt.org/#/q/topic:integration [2] https://gerrit.ovirt.org/#/c/47008/17 [3] https://gerrit.ovirt.org/#/c/47007/10 [4] https://gerrit.ovirt.org/#/c/47008/17 [5] https://maven.apache.org/surefire/maven-failsafe-plugin/

5 7

drop-in directories for vdsm configuration
by Irit Goihman 04 Jul '16

04 Jul '16

Hi, A drop-in directories feature has been added to vdsm-4.18.4. This allows the user to put partial configurations into a drop-in file and create new parameters or override default ones. The semantics of the directories is as follows: - /etc/vdsm/vdsm.conf: For default configuration. We install this file if missing, and never touch this file during upgrade. - /etc/vdsm/vdsm.conf.d/ - / run/vdsm/vdsm.conf.d/ - /usr/lib/vdsm/vdsm.conf.d/ Files with a .conf suffix can be placed into any of the vdsm.conf.d drop-in directories. The default configuration file is read before all other files and has the lowest priority. The priority of configuration files under vdsm.conf.d is determined by lexicographic order of file names, so the lexicographic latest file name takes precedence. It is highly recommended to prefix all file names with a two-digit number in order to keep the ordering simple. -- Irit Goihman Software Engineer Red Hat Israel Ltd.

1 0