- Devel - oVirt List Archives

VDSM have moved to github
by Nir Soffer 09 Jan '22

09 Jan '22

We are not using the gerrit project now, and it was changed to read-only. We are accepting now pull requests on github: https://github.com/oVirt/vdsm/pulls To contribute, use the standard github flow: 1. Fork vdsm on github Example: https://github.com/nirs/vdsm (you need to create a github account if you don't have one) 2. Clone your fork Example: git clone git@github.com:nirs/vdsm.git 3. Push changes to your fork and submit pull requests Example: https://github.com/oVirt/vdsm/pull/28 4. Check the github CI for your pull request Example: https://github.com/oVirt/vdsm/pull/28/checks Note: We don't do squash merges. When addressing comments, force push to your fork. Example: https://github.com/oVirt/ovirt-imageio/pull/15 Common issues: 1. How do I add reviewers to my pull request? Once https://github.com/oVirt/vdsm/pull/28 will be merged, we will have a CODEOWNERS file: https://github.com/oVirt/vdsm/blob/96320682a8b41e440fbbb8ca624fd73efa95bda0… so github will assign reviews automatically. Regardless you can add a comment on you patch like: @{reviewer} please review. 2. How do I run OST with my pull request? OST does not run yet automatically. You have 2 choices: - Run OST locally (ask here how to run it) - Use the Red Hat OST runner (ask on rhev-devel(a)redhat.com how to use it) When running OST, you can use the built rpms from your pull request: https://github.com/oVirt/vdsm/actions/runs/1673718827 (see the artifacts.zip at the bottom) 3. How to backport a patch to stable branches? Backport your patch(es) on top of the right stable branch (e.g. ovirt-4.4.z). When you create a pull request, please choose the right branch. The stable branches maintainers will be added for review automatically. 4. How to compare changes when a contributor submitted a new version? When contributor force pushes, github shows: nirs *force-pushed* the inactivity-timeout branch 2 times, most recently from *790cc4c* to *38d08c8* 22 days ago Click the "force-pushed" link to see the differences: https://github.com/oVirt/ovirt-imageio/compare/790cc4c44c9c6577dae0badb3fc6… 5. How to download a pull request to my machine? You can use the github command lline tool "gh": gh pr checkout 28 See https://cli.github.com/ for more info. (note: I did not try it yet) 6. How do I comment on commit message? Github does not have this feature, you can use the review comment, or add a separate comment. 7. How to run CI on a single patch in my PR? Github runs the CI for the entire PR. If your PR have several commits, issues in early commits fixed in later commits are hidden. To verify that all commits pass the tests you can post the submit a new branch to your local fork. Example flow: git checkout -b testing-foobar commit-id git push origin testing-foobar (check CI on your fork) Alternatively, use smaller pull requests. You can submit pull request based on another pull request. Nir

1 0

Need help in dumping a table
by Matthew.Stier＠fujitsu.com 05 Jan '22

05 Jan '22

I'm planning a re-organization of my oVirt environment. I need a Postgres command-line to dump a listing of VMs/disks/storage_domains.

1 0

Cannot build Engine rpm's on CentOS Stream 8
by Milan Zamazal 05 Jan '22

05 Jan '22

Hi, I've been having trouble to build Engine rpm's locally on up-to-date Stream 8 since about mid December, most likely due to javapackages-local update to 6.0.0-1.4.el8s. Engine compiles fine but building the rpm's fails on missing %add_maven_depmap rpm spec macro. There is a patch for Stream 9: https://gerrit.ovirt.org/c/ovirt-engine/+/116811. But when I rebase on it, I get the following error: [INFO maven.req] input: ['/root/rpmbuild/BUILDROOT/ovirt-engine-4.5.0-0.2.master.20220104124402.el8.x86_64/usr/share/maven-metadata/ovirt-engine-base.xml'] Traceback (most recent call last): File "/usr/lib/rpm/maven.req", line 393, in <module> builder = TagBuilder() File "/usr/lib/rpm/maven.req", line 71, in __init__ self.print_requires(metadata) File "/usr/lib/rpm/maven.req", line 200, in print_requires self._handle_unknown_deps(unknown_deps) File "/usr/lib/rpm/maven.req", line 258, in _handle_unknown_deps raise Exception(unknown_msg) Exception: Following dependencies were not resolved and requires cannot be generated. Either remove the dependencies from pom.xml or add proper packages to BuildRequires: commons-beanutils:commons-beanutils:1.9.4 required by org.ovirt.engine.core:utils commons-lang:commons-lang:2.6 required by org.ovirt.engine.core:aaa org.infinispan:infinispan-core:9.4.18.Final required by org.ovirt.engine.core:utils commons-collections:commons-collections:3.2.2 required by org.ovirt.engine.core:bll org.jboss.modules:jboss-modules:1.1.1.GA required by org.ovirt.engine.core:common org.hibernate.validator:hibernate-validator:6.0.18.Final required by org.ovirt.engine.core:common This happens with both maven 3.5 and 3.6. Any idea what's wrong and how to fix it on Stream 8? Thanks, Milan

2 3

Re: [oVirt/ovirt-release] Failed repoclosure job (Issue #60)
by Yedidyah Bar David 05 Jan '22

05 Jan '22

On Wed, Jan 5, 2022 at 1:09 PM github-actions[bot] <notifications(a)github.com> wrote: > ❌ The repoclosure CI job is still failing. Please investigate. > <https://github.com/oVirt/ovirt-release/actions/runs/1658041750> > It failed because we recently started building metrics and dwh on el9, which require stuff we do not have: - The engine, already discussed elsewhere, patches are WIP (apparently waiting for Artur? Not sure). - ansible. No idea why it's missing. Perhaps better not build new stuff on el9 before we fix these, or stop running repoclosure on el9 for now. Best regards, -- Didi

2 2

create a host using ssh
by Grace Chen 03 Jan '22

03 Jan '22

I am writing a script to create a host using public ssh key the function I am using is add_using_ssh: my ssh_key is a an Ssh type already ssh_key = types.Ssh(public_key=my_ssh_key_string) host = hosts_service.add_using_ssh( types.Host( name=name, description='', address=address, root_password=None, port=22, ssh=ssh_key, cluster=types.Cluster( name=cluster, ), ), ) When I choose from GUI, it doesn't ask me to put in root password, so I set it as None (Not sure if this is correct) I got error: ovirtsdk4.Error: Fault reason is "Request syntactically incorrect.". Fault detail is "For correct usage, see: ovirt-engine/apidoc#services/hosts/methods/add". HTTP response code is 400. looks like I didn't get all the parameters assigned? Can anybody help with this?

2 5

Re: [oVirt/ovirt-release] Failed repoclosure job (Issue #55)
by Yedidyah Bar David 28 Dec '21

28 Dec '21

On Mon, Dec 27, 2021 at 2:38 PM github-actions[bot] <notifications(a)github.com> wrote: > > ❌ The repoclosure CI job is still failing. Please investigate. I think this started failing recently (on el9) due to a combination of: - The engine is still not built on el9 in copr. I tried building it myself there and so far got stuck - it seems like the fix is an existing pending patch [1] (failed at %add_maven_depmap). - ovirt-engine-keycloak is now built also for el9, but requires an engine package, so we fail on repoclosure. Unless we have concrete plans to fix this very soon, I suggest to stop the noise for now - either not build (or publish?) keycloak for el9, or do not run repoclosure on el9, or perhaps filter the output to not fail if it's only on these packages, etc. [1] https://gerrit.ovirt.org/c/ovirt-engine/+/116811 -- Didi

2 2

OST fails with "Error loading module from /usr/share/ovirt-engine/modules/common/org/springframework/main/module.xml"
by Marcin Sobczyk 27 Dec '21

27 Dec '21

Hi All, OST currently fails all the time during engine setup. Here's a piece of ansible log that's seen repeatedly and I think describes the problem: 11:07:54 E "engine-config", 11:07:54 E "-s", 11:07:54 E "OvfUpdateIntervalInMinutes=10" 11:07:54 E ], 11:07:54 E "delta": "0:00:01.142926", 11:07:54 E "end": "2021-12-24 11:06:37.894810", 11:07:54 E "invocation": { 11:07:54 E "module_args": { 11:07:54 E "_raw_params": "engine-config -s OvfUpdateIntervalInMinutes='10' ", 11:07:54 E "_uses_shell": false, 11:07:54 E "argv": null, 11:07:54 E "chdir": null, 11:07:54 E "creates": null, 11:07:54 E "executable": null, 11:07:54 E "removes": null, 11:07:54 E "stdin": null, 11:07:54 E "stdin_add_newline": true, 11:07:54 E "strip_empty_ends": true, 11:07:54 E "warn": false 11:07:54 E } 11:07:54 E }, 11:07:54 E "item": { 11:07:54 E "key": "OvfUpdateIntervalInMinutes", 11:07:54 E "value": "10" 11:07:54 E }, 11:07:54 E "msg": "non-zero return code", 11:07:54 E "rc": 1, 11:07:54 E "start": "2021-12-24 11:06:36.751884", 11:07:54 E "stderr": "Picked up JAVA_TOOL_OPTIONS: -Dcom.redhat.fips=false", 11:07:54 E "stderr_lines": [ 11:07:54 E "Picked up JAVA_TOOL_OPTIONS: -Dcom.redhat.fips=false" 11:07:54 E ], 11:07:54 E "stdout": "Error loading module from /usr/share/ovirt-engine/modules/common/org/springframework/main/module.xml", 11:07:54 E "stdout_lines": [ 11:07:54 E "Error loading module from /usr/share/ovirt-engine/modules/common/org/springframework/main/module.xml" We do set some config values for the engine in OST when running engine-setup. I tried commenting these out, but then engine failed health check anyway: "Status code was 503 and not [200]: HTTP Error 503: Service Unavailable" The last working set of OST images was the one from Dec 23, 2021 2:05:08 AM. The first broken one is from Dec 24, 2021 2:05:09 AM. The shipped ovirt-engine RPMs doesn't seem to contain any important changes for these two sets, but AFAICS the newer ovirt-dependencies RPM did take in a couple of patches that look suspicious [1][2][3]. The patches were merged on November 16th, but it seems they were first used in that broken set from Dec 24 (the one from Dec 23 seems to contain ovirt-dependencies RPM based on this [4] commit). I wanted to try out an older version of ovirt-dependencies, but I think they were wiped out from resources.ovirt.org. I will disable cyclic el8stream OST runs for now, cause all of them fail. If there's anyone working and able to make a build with those patches reverted and test that out, please ping me and I'll re-enable them. Regards, Marcin [1] https://gerrit.ovirt.org/c/ovirt-dependencies/+/114699 [2] https://gerrit.ovirt.org/c/ovirt-dependencies/+/113877 [3] https://gerrit.ovirt.org/c/ovirt-dependencies/+/114654 [4] https://gerrit.ovirt.org/c/ovirt-dependencies/+/117459

2 1

Re: [ovirt-users] Re: using stop_reason as a vdsm hook trigger into the UI
by Nathanaël Blanchet 22 Dec '21

22 Dec '21

Le 22/12/2021 à 14:56, John Taylor a écrit : > Maybe use the events api and search for the shutdown and reason in there? > > api/events;from={event_id}?search={query}" rel="events/search"/> > > -John Exact! that's precisely the workaround I already tested this morning as well :) and it works because the event contains the stop_reason in the description field before vm is in the stopped status! > > On Tue, Dec 21, 2021 at 9:00 AM Nathanaël Blanchet <blanchet(a)abes.fr > <mailto:blanchet@abes.fr>> wrote: > > Thanks for responding, > > Le 20/12/2021 à 21:42, Nir Soffer a écrit : >> On Mon, Dec 20, 2021 at 9:59 PM Nathanaël Blanchet<blanchet(a)abes.fr> <mailto:blanchet@abes.fr> wrote: >> >> Adding the devel list since question is more about extending oVirt >> ... >>> The idea is to use the stop_reason element into the vm xml definition. But after hours, I realized that this element is writed to the vm definition file only after the VM has been destroyed. >> So you want to run the clean hook only if stop reason == "clean"? >> >> I think the way to integrate hooks is to define a custom property >> in the vm, and check if the property was defined in the hook. >> >> For example how the localdisk hook is triggered: >> >> def main(): >> backend = os.environ.get('localdisk') >> if backend is None: >> return >> if backend not in [BACKEND_LVM, BACKEND_LVM_THIN]: >> hooking.log("localdisk-hook: unsupported backend: %r" % backend) >> return >> ... >> >> The hook runs only if the environment variable "localdisk" is defined >> and configured properly. >> >> vdsm defines the custom properties as environment variables. >> >> On the engine side, you need to add a user defined property: >> >> engine-config -s UserDefinedVMProperties='localdisk=^(lvm|lvmthin)$' >> >> And configure a custom property with one of the allowed values, like: >> >> localdisk=lvm >> >> See vdsm_hooks/localdisk/README for more info. >> >> If you want to control the cleanup, by adding a "clean" stop reason only when >> needed, this will not help, and vdsm hook is probably not the right way >> to integrate this. > Sure >> If your intent is to clean a vm in some special events, but you want >> to integrate >> this in engine, maybe you should write an engine ui plugin? >> >> The plugin can show the running vms, and provide a clean button that will >> shut down the vm and run your custom code. > too complex for doing what I want >> But maybe you don't need to integrate this in engine, and having a simple >> script using ovirt engine API/SDK to shutdown the vm and run the cleanup >> code. > My playbook/scripts work already fine, but this is not my goal. >> Nir >> > I will sum up my initial question: *Is there any way to get the > value of "stop_reason" (value of the field in the UI) so as to > reuse this variable into a vdsm hook?* > > Thank you > > > -- > Nathanaël Blanchet > > Supervision réseau > SIRE > 227 avenue Professeur-Jean-Louis-Viala > 34193 MONTPELLIER CEDEX 5 > Tél. 33 (0)4 67 54 84 55 > Fax 33 (0)4 67 54 84 14 > blanchet(a)abes.fr <mailto:blanchet@abes.fr> > > _______________________________________________ > Users mailing list -- users(a)ovirt.org <mailto:users@ovirt.org> > To unsubscribe send an email to users-leave(a)ovirt.org > <mailto:users-leave@ovirt.org> > Privacy Statement: https://www.ovirt.org/privacy-policy.html > <https://www.ovirt.org/privacy-policy.html> > oVirt Code of Conduct: > https://www.ovirt.org/community/about/community-guidelines/ > <https://www.ovirt.org/community/about/community-guidelines/> > List Archives: > https://lists.ovirt.org/archives/list/users@ovirt.org/message/OSM572SLKKAFO… > <https://lists.ovirt.org/archives/list/users@ovirt.org/message/OSM572SLKKAFO…> > -- Nathanaël Blanchet Supervision réseau SIRE 227 avenue Professeur-Jean-Louis-Viala 34193 MONTPELLIER CEDEX 5 Tél. 33 (0)4 67 54 84 55 Fax 33 (0)4 67 54 84 14 blanchet(a)abes.fr

1 1

Re: [ovirt-users] using stop_reason as a vdsm hook trigger into the UI
by Nir Soffer 21 Dec '21

21 Dec '21

On Mon, Dec 20, 2021 at 9:59 PM Nathanaël Blanchet <blanchet(a)abes.fr> wrote: Adding the devel list since question is more about extending oVirt ... > The idea is to use the stop_reason element into the vm xml definition. But after hours, I realized that this element is writed to the vm definition file only after the VM has been destroyed. So you want to run the clean hook only if stop reason == "clean"? I think the way to integrate hooks is to define a custom property in the vm, and check if the property was defined in the hook. For example how the localdisk hook is triggered: def main(): backend = os.environ.get('localdisk') if backend is None: return if backend not in [BACKEND_LVM, BACKEND_LVM_THIN]: hooking.log("localdisk-hook: unsupported backend: %r" % backend) return ... The hook runs only if the environment variable "localdisk" is defined and configured properly. vdsm defines the custom properties as environment variables. On the engine side, you need to add a user defined property: engine-config -s UserDefinedVMProperties='localdisk=^(lvm|lvmthin)$' And configure a custom property with one of the allowed values, like: localdisk=lvm See vdsm_hooks/localdisk/README for more info. If you want to control the cleanup, by adding a "clean" stop reason only when needed, this will not help, and vdsm hook is probably not the right way to integrate this. If your intent is to clean a vm in some special events, but you want to integrate this in engine, maybe you should write an engine ui plugin? The plugin can show the running vms, and provide a clean button that will shut down the vm and run your custom code. But maybe you don't need to integrate this in engine, and having a simple script using ovirt engine API/SDK to shutdown the vm and run the cleanup code. Nir

2 1

selinux for Ganesha absent in master?
by lejeczek 21 Dec '21

21 Dec '21

Hi guys. I get: -> $ dnf install glusterfs-ganesha.x86_64 Mine and private 2.9 MB/s | 3.0 kB 00:00 Error: Problem: cannot install the best candidate for the job - nothing provides nfs-ganesha-gluster >= 2.7.6 needed by glusterfs-ganesha-10.0-1.el9s.x86_64 that is with: ovirt-release-master-4.5.0-0.0.master.20211206152702.gitebb0229.el9.noarch many thanks, L.

3 2