- Users - oVirt List Archives

Localdisk hook not working
by tim-nospam＠bordemann.com 30 Jun '20

30 Jun '20

Hi, I'd like to use the localdisk hook of vdsm and have configured everything according to the readme: https://github.com/oVirt/vdsm/tree/master/vdsm_hooks/localdisk After installing the hook, configuring the ovirt-engine, creating the volume group, adding the custom property 'localdisk' to the virtual machine and fixing a small bug in the localdisk-helper, vdsm creates the logical volume on the 'ovirt-local' volume group after starting the virtual machine. Unfortunately nothing more seems to happen then. There is no activity on the NAS which would indicate that the disk image is pulled to the host. I can also see no errors on the host. Is anyone currently running ovirt 4.4 with the localdisk hook? What else can I do to find out why the image is not being copied to my host? Thanks, Tim

2 2

Re: VDSM not binding on IPv4 after ovirt-engine restart
by Dominik Holler 30 Jun '20

30 Jun '20

On Tue, Jun 30, 2020 at 6:13 PM Erez Zarum <erezz(a)nanosek.com> wrote: > While troubleshooting a fresh installation of (after a failed one) that > caused all the hosts but the one running the hosted-engine to become in > “Unassigned” state I noticed that the ovirt-engine complains about not > being able to contact the VDSM. > I noticed that VDSM has stopped listening on IPv4. > > Thanks for sharing the details. > I didn’t disable any IPv6 as it states not to disable it on hosts that are > capable running the hosted-engine and it seems that the reason behind it > is that the hosted-engine talks to the host it runs on through “localhost”, > this also explains why the host which the hosted-engine runs on is “OK”. > > Below is from a host that does not run the hosted-engine: > # ss -atn | grep 543 > LISTEN 0 5 *:54322 *:* > ESTAB 0 0 127.0.0.1:54792 127.0.0.1:54321 > ESTAB 0 0 127.0.0.1:54798 127.0.0.1:54321 > LISTEN 0 5 [::]:54321 [::]:* > ESTAB 0 0 [::ffff:127.0.0.1]:54321 > [::ffff:127.0.0.1]:54798 > ESTAB 0 0 [::ffff:127.0.0.1]:54321 > [::ffff:127.0.0.1]:54792 > ESTAB 0 0 [::1]:54321 [::1]:50238 > ESTAB 0 0 [::1]:50238 [::1]:54321 > > Below is from a host that runs the hosted-engine at the moment: > # ss -atn | grep 543 > LISTEN 0 5 *:54322 *:* > LISTEN 0 5 [::]:54321 [::]:* > ESTAB 0 0 [::1]:51230 [::1]:54321 > ESTAB 0 0 [::1]:54321 [::1]:51242 > ESTAB 0 0 [::ffff:10.46.20.23]:54321 > [::ffff:10.46.20.20]:45706 > ESTAB 0 0 [::ffff:10.46.20.23]:54321 > [::ffff:10.46.20.20]:45746 > ESTAB 0 0 [::1]:51240 [::1]:54321 > ESTAB 0 0 [::1]:54321 [::1]:51230 > ESTAB 0 0 [::1]:51242 [::1]:54321 > ESTAB 0 0 [::1]:54321 [::1]:51240 > > The hosted-engine IP is 10.46.20.20 and the host is 10.46.20.23. > > Why do you think the host does not listen to IPv4 anymore? Can you please share the output of "nc -vz 10.46.20.23 54321" executed on engine VM or another host? > /etc/hosts on all hosts: > 127.0.0.1 localhost localhost.localdomain localhost4 > localhost4.localdomain4 > ::1 localhost localhost.localdomain localhost6 > localhost6.localdomain6 > > Perhaps this is relevant but all hosts are enrolled into IDM (FreeIPA) and > as an outcome they all have a DNS record and a PTR record as well as the > ovirt-engine VM. > > # cat /etc/vdsm/vdsm.conf > [vars] > ssl = true > ssl_ciphers = HIGH:!aNULL > ssl_excludes = OP_NO_TLSv1,OP_NO_TLSv1_1 > > [addresses] > management_port = 54321 > > I have tried adding “management_ip = 0.0.0.0” but then it only binds to > IPv4 and yet, the host still shows as Unassigned, sometimes it switches to > “NonResponsive” and trying to “Reinstall” the host fails, the ovirt-engine > complains it can't contact/reach the VDSM, while using netcat from the > ovirt-engine it works. > > I have KSM and Memory Ballooning enabled on the Cluster as well. > > oVirt 4.3.10 installed on CentOS 7.8.2003 > The self-hosted Engine runs on an external GlusterFS, before reinstalling > everything (fresh start of OS, etc..) I tried iSCSI as well. > > > _______________________________________________ > Users mailing list -- users(a)ovirt.org > To unsubscribe send an email to users-leave(a)ovirt.org > Privacy Statement: https://www.ovirt.org/privacy-policy.html > oVirt Code of Conduct: > https://www.ovirt.org/community/about/community-guidelines/ > List Archives: > https://lists.ovirt.org/archives/list/users@ovirt.org/message/NCTWZLS2VPIAB… >

2 1

VDSM not binding on IPv4 after ovirt-engine restart
by Erez Zarum 30 Jun '20

30 Jun '20

While troubleshooting a fresh installation of (after a failed one) that caused all the hosts but the one running the hosted-engine to become in “Unassigned” state I noticed that the ovirt-engine complains about not being able to contact the VDSM. I noticed that VDSM has stopped listening on IPv4. I didn’t disable any IPv6 as it states not to disable it on hosts that are capable running the hosted-engine and it seems that the reason behind it is that the hosted-engine talks to the host it runs on through “localhost”, this also explains why the host which the hosted-engine runs on is “OK”. Below is from a host that does not run the hosted-engine: # ss -atn | grep 543 LISTEN 0 5 *:54322 *:* ESTAB 0 0 127.0.0.1:54792 127.0.0.1:54321 ESTAB 0 0 127.0.0.1:54798 127.0.0.1:54321 LISTEN 0 5 [::]:54321 [::]:* ESTAB 0 0 [::ffff:127.0.0.1]:54321 [::ffff:127.0.0.1]:54798 ESTAB 0 0 [::ffff:127.0.0.1]:54321 [::ffff:127.0.0.1]:54792 ESTAB 0 0 [::1]:54321 [::1]:50238 ESTAB 0 0 [::1]:50238 [::1]:54321 Below is from a host that runs the hosted-engine at the moment: # ss -atn | grep 543 LISTEN 0 5 *:54322 *:* LISTEN 0 5 [::]:54321 [::]:* ESTAB 0 0 [::1]:51230 [::1]:54321 ESTAB 0 0 [::1]:54321 [::1]:51242 ESTAB 0 0 [::ffff:10.46.20.23]:54321 [::ffff:10.46.20.20]:45706 ESTAB 0 0 [::ffff:10.46.20.23]:54321 [::ffff:10.46.20.20]:45746 ESTAB 0 0 [::1]:51240 [::1]:54321 ESTAB 0 0 [::1]:54321 [::1]:51230 ESTAB 0 0 [::1]:51242 [::1]:54321 ESTAB 0 0 [::1]:54321 [::1]:51240 The hosted-engine IP is 10.46.20.20 and the host is 10.46.20.23. /etc/hosts on all hosts: 127.0.0.1 localhost localhost.localdomain localhost4 localhost4.localdomain4 ::1 localhost localhost.localdomain localhost6 localhost6.localdomain6 Perhaps this is relevant but all hosts are enrolled into IDM (FreeIPA) and as an outcome they all have a DNS record and a PTR record as well as the ovirt-engine VM. # cat /etc/vdsm/vdsm.conf [vars] ssl = true ssl_ciphers = HIGH:!aNULL ssl_excludes = OP_NO_TLSv1,OP_NO_TLSv1_1 [addresses] management_port = 54321 I have tried adding “management_ip = 0.0.0.0” but then it only binds to IPv4 and yet, the host still shows as Unassigned, sometimes it switches to “NonResponsive” and trying to “Reinstall” the host fails, the ovirt-engine complains it can't contact/reach the VDSM, while using netcat from the ovirt-engine it works. I have KSM and Memory Ballooning enabled on the Cluster as well. oVirt 4.3.10 installed on CentOS 7.8.2003 The self-hosted Engine runs on an external GlusterFS, before reinstalling everything (fresh start of OS, etc..) I tried iSCSI as well.

1 0

4.4.1-rc5: Looking for correct way to configure machine=q35 instead of machine=pc for arch=x86_64
by Glenn Marcy 29 Jun '20

29 Jun '20

Hello, I am hoping for some insight from folks with more hosted engine install experience. When I try to install the hosted engine using the RC5 dist I get the following error during the startup of the HostedEngine VM: XML error: The PCI controller with index='0' must be model='pci-root' for this machine type, but model='pcie-root' was found instead This is due to the HE Domain XML description using machine="pc-i440fx-rhel7.6.0". I've tried to override the default of 'pc' from ovirt-ansible-hosted-engine-setup/defaults/main.yml: he_emulated_machine: pc by passing to the ovirt-hosted-engine-setup script a --config-append=file parameter where file contains: [environment:default] OVEHOSTED_VM/emulatedMachine=str:q35 When the "Create ovirt-hosted-engine-ha run directory" step finishes the vm.conf file contains: cpuType=IvyBridge,+pcid,+spec-ctrl,+ssbd,+md-clear emulatedMachine=q35 At the "Start ovirt-ha-broker service on the host" step that file is removed. When that file appears again during the "Check engine VM health" step it now contains: cpuType=IvyBridge,+pcid,+spec-ctrl,+ssbd,+md-clear emulatedMachine=pc-i440fx-rhel7.6.0 After that the install fails with the metadata from "virsh dumpxml HostedEngine" containing: <ovirt-vm:exitCode type="int">1</ovirt-vm:exitCode> <ovirt-vm:exitMessage>XML error: The PCI controller with index='0' must be model='pci-root' for this machine type, but model='pcie-root' was found instead</ovirt-vm:exitMessage> Interestingly enough, the HostedEngineLocal VM that is running the appliance image has the value I need: <type arch='x86_64' machine='pc-q35-rhel8.2.0'>hvm</type> Does anyone on the list have any experience with where this needs to be overridden? Somewhere in the hosted engine setup or do I need to do something at a deeper level like vdsm or libvirt? Help much appreciated ! Thanks, Glenn

2 3

vdsm error every hour
by eevans＠digitaldatatechs.com 29 Jun '20

29 Jun '20

In the web ui I get an error about every hour: VDSM command SetVolumeDescriptionVDS failed: Volume does not exist: (u'e3f79840-8355-45b0-ad2b-440c877be637',) I looked in storage and disks and this disk does not exist. Its more of an annoyance than a problem but if there is way to get rid of this error I would like to know. My research says I can install vdsm-tools and vdsm-cli but vdsm-cli is not available and I really don;t want to install anything until I know it's what I need. Is there a vdsm command to purge a missing disk so this error won't show up? Thanks in advance. Eric

2 1

Error with Imported VM Run
by Ian Easter 29 Jun '20

29 Jun '20

I imported an orphaned Storage Domain into a new oVirt instance (4.4) and imported the VMs that were found in that storage domain. The SD was previously attached to, 4.2/4.3 instance of oVirt and I did not see any errors about the imports. However, when I attempt to run these VMs, I see this error in the engine.log: ERROR [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] (EE-ManagedThreadFactory-engine-Thread-51848) [d56a3ce0-7139-4173-bcdb-d5bd4d9378fa] EVENT_ID: USER_FAILED_RUN_VM(54), Failed to run VM mtl-portal-01 (User: admin@internal-authz). As best as I can tell, this is a vdsm version issue... So what are my options to resolve this? Should I go through the steps to build out an older oVirt release and then attach the SD to that instance and work through the migration? Can I manually update the VM to work with the newer vdsm version? *Thank you,* *Ian Easter*

2 2

NVMe-oF
by Daniel Menzel 29 Jun '20

29 Jun '20

Hi all, does anybody know whether NVMe-over-Fabrics support is somehow scheduled in oVirt? As, from a logical point of view, it is like iSCSI on steroids, I guess it shouldn't be too hard to implement it. Yep, I know, bold statement from someone who isn't a programmer. ;-) Nonetheless: Are there plans to implement NVMe-oF as storage backend for oVirt in the near future? If so: Is there a way to help (i.e. with hardware ressources)? Kind regards, Daniel -- Daniel Menzel Geschäftsführer Menzel IT GmbH Charlottenburger Str. 33a 13086 Berlin +49 (0) 30 / 5130 444 - 00 daniel.menzel(a)menzel-it.net https://menzel-it.net Geschäftsführer: Daniel Menzel, Josefin Menzel Unternehmenssitz: Berlin Handelsregister: Amtsgericht Charlottenburg Handelsregister-Nummer: HRB 149835 B USt-ID: DE 309 226 751

2 2

oVirt 2020 online conference
by Sandro Bonazzola 29 Jun '20

29 Jun '20

It is our pleasure to invite you to oVirt 2020 online conference. The conference,organized by oVirt community, will take place online on Monday, September 7th 2020! oVirt 2020 is a free conference for oVirt community project users and contributors coming to a web browser near you! There is no admission or ticket charge for this event. However, you will be required to complete a free registration. Watch https://blogs.ovirt.org/ovirt-2020-online-conference/ for updates about registration. Talks, presentations and workshops will all be in English. We encourage students and new graduates as well as professionals to submit proposals to oVirt conferences. We will be looking for talks and discussions across virtualization, and how oVirt 4.4 can effectively solve user issues around: - Upgrade flows - New features - Integration with other projects - User stories The deadline to submit abstracts is July 26th 2020. To submit your abstract, please click on the following link: submission form <https://forms.gle/1RDUwPGKobJfKhaYA> More information are available at https://blogs.ovirt.org/ovirt-2020-online-conference/ Thanks, -- Sandro Bonazzola MANAGER, SOFTWARE ENGINEERING, EMEA R&D RHV Red Hat EMEA <https://www.redhat.com/> sbonazzo(a)redhat.com <https://www.redhat.com/> *Red Hat respects your work life balance. Therefore there is no need to answer this email out of your office hours. <https://mojo.redhat.com/docs/DOC-1199578>*

1 0

Re: VDSM can't see StoragePool
by Strahil Nikolov 28 Jun '20

28 Jun '20

Can you set one of the Hypervisours into maintenance and use the "reinstall" option from the UI ? Best Regards, Strahil Nikolov На 25 юни 2020 г. 13:24:26 GMT+03:00, Erez Zarum <erezz(a)nanosek.com> написа: >I have a Self-hosted Engine running on iSCSI as well as couple of >Storage domains using iSCSI, both the SE and those Storage Domains uses >the same target portals (two). >I can see the iSCSI sessions and multipath working well from the Host >point of view. >Yesterday after doing a restart for the “ovirt-engine” all the hosts >besides the one that runs the SE and the SPM went into “Unassigned” >mode with an error stating the ovirt-engine can’t communicate with the >hosts. >Network wise, everything is good, I can reach all the ports, all the >network is well configured, so I ruled this. >Looking at the VDSM logs on those “Unassigned” hosts it looks like the >VDSM can’t find the Storage Pool. > >(vmrecovery) [vdsm.api] START >getConnectedStoragePoolsList(options=None) from=internal, >task_id=217ec32b-591c-4376-8dc0-8d62200557ee (api:48) >(vmrecovery) [vdsm.api] FINISH getConnectedStoragePoolsList >return={'poollist': []} from=internal, >task_id=217ec32b-591c-4376-8dc0-8d62200557ee (api:54) >(vmrecovery) [vds] recovery: waiting for storage pool to go up >(clientIF:723) > >If I look at the VDSM logs on the host where the SE and SPM is running, >no issues there and the node appears up (green) in the ovirt-engine UI. > >I managed to set the hosted-engine to maintenance, shut it down and >then start it again on another Host, when it starts on that Host, the >host goes “green” and if the SPM stays on the previous host, I have two >hosts working and the rest remains “Unassigned”. > >All the “ovirt-ha-agent”/”ovirt-ha-broker” services seems ok, I >restarted them, I also tried to restart the VDSM on the hosts with no >luck. >I have the VMs still running, I did shutdown one host (even used the >“SSH restart” from the WebUI) to see if that helps, it came back and >still went into “Unassigned”. > >It seems like the hosts can’t see the Storage pool. > >Where should I start to troubleshoot this? > >Thanks

1 0

Re: Migration of self Hosted Engine from iSCSI to Gluster/NFS
by Strahil Nikolov 28 Jun '20

28 Jun '20

As you will migrate from block-based storage to file-based storage, I think that you should use the backup & restore procedure. Best Regards, Strahil Nikolov На 25 юни 2020 г. 7:31:55 GMT+03:00, Erez Zarum <erezz(a)nanosek.com> написа: >I was looking for a “complete” best practice to migrate a self-hosted >engine running currently on an iSCSI LUN to a Gluster of NFS storage >domain >oVirt version 4.3.10 > >Thanks!

1 0