- Users - oVirt List Archives

Network Performance Issue on oVirt Node Using Intel Platinum 8562Y+ (Sapphire Rapids) CPU
by dhayanands＠botvfx.com 17 Jun '25

17 Jun '25

Hi team, We’re facing a strange issue with network performance on one of our oVirt hosts and would like to get insights or guidance on possible misconfigurations or compatibility problems. Environment Details: oVirt Version: 4.5.5 Cluster CPU Type: Intel IceLake Server Family (default) Host A (Issue observed): CPU: Intel Xeon Platinum 8562Y+ (Sapphire Rapids) CPU Family/Model: Family 6 / Model 207 / Stepping 2 Host B (Working fine): CPU: Intel Xeon Gold 6354 (Ice Lake) CPU Family/Model: Family 6 / Model 106 / Stepping 6 Problem Description: We have a VM (kitsu) that was initially hosted on botovirtnode-006 (Platinum 8562Y+). The VM showed very poor download speeds (~10 Mbps), while upload was fine (~680 Mbps). After migrating the same VM to another host (botovirtnode-001 with Gold 6354 CPU), network performance drastically improved: Test On Affected Node On Healthy Node Download ~10 Mbps ~920 Mbps Upload ~680 Mbps ~760 Mbps This clearly indicates a host-specific issue, most likely related to hardware compatibility or CPU feature masking. Troubleshooting Done: Verified VM configurations (same on both hosts). Tested with speedtest-cli multiple times to confirm. Confirmed that VM virtual NIC, vNIC profile, and vSwitch settings are identical. Both hosts are part of the same cluster using the default Intel IceLake Server CPU type. Suspect the issue is due to CPU feature mismatch or unsupported stepping/model in the cluster’s CPU type.

3 3

Re: Network Performance Issue on oVirt Node Using Intel Platinum 8562Y+ (Sapphire Rapids) CPU
by Domien Van Rompaey 16 Jun '25

16 Jun '25

Hi there I was just about to post a similar issue when I saw your post. We have a cluster of 3 AMD EPYC 9454 48-Core Processor hosts who have BCM57504 NICs in a LACP bond to 2 Dell S5248F-ON switches. We are seeing assymetrical network performance between VM's making the cluster unusable. In previous test setups we did not have this issue. The issues seems to manifest only for traffic between hosts over VLAN networks and SDNs do not seem to be affected. If we do an upload test over a virtual firewall we get the same numbers as you do. If we do some iperf tests we get this: (We will use the name VM1 and VM2 who are in the same VLAN network. For hosts we will use Host1 and Host2 in the names) VM1: 10.22.21.10 VM2: 10.22.21.20 VM1 (Host2) to VM2 (Host1) [cid:image002.png@01DBDEA1.6BFA2C60] VM2 (Host1) -> VM1 (Host2) [cid:image003.png@01DBDEA1.6BFA2C60] When we put these on the same host (Host2) VM2 -> VM1 [cid:image004.png@01DBDEA1.6BFA2C60] VM1->VM2 [cid:image005.png@01DBDEA1.6BFA2C60] I’m believing the underlying issue will be the same. If it’s not I will probably split my issue in a seperate post. We raised a ticket with Oracle Support for this (we are running OLVM, latest version) but it’s going very slow. The issue is stopping us from bringing our cluster into production. Met vriendelijke groeten, Domien Van Rompaey Security Engineer Cloud & ICT‑infrastructure | Security Tel +32 51 76 00 00 [https://website-api.dynamate.be/public/E-mailhandtekening/DYNAMATE-handteke…]<https://eu.content.exclaimer.net/?url=https%3A%2F%2Fwww.christiaens.net%2F&…> www.dynamate.be<https://eu.content.exclaimer.net?url=https%3A%2F%2Fwww.dynamate.be%2F&tenan…> [https://website-api.dynamate.be/public/E-mailhandtekening/instagram-60-150.…]<https://eu.content.exclaimer.net/?url=https%3A%2F%2Fwww.instagram.com%2Fwea…> [https://website-api.dynamate.be/public/E-mailhandtekening/facebook-60-150.p…]<https://eu.content.exclaimer.net/?url=https%3A%2F%2Fwww.facebook.com%2Fwear…> [https://website-api.dynamate.be/public/E-mailhandtekening/linkedin-60-150-2…]<https://eu.content.exclaimer.net/?url=https%3A%2F%2Fwww.linkedin.com%2Fcomp…>

1 0

Unable to Start the Hosted Engine and the WebUI
by kelwintai＠proton.me 13 Jun '25

13 Jun '25

Hi there, I have tried to restart the Hosted Engine but I cannot ssh/ping the Hosted Engine after the restart. The oVirt webUI was down and I cannot start the VM in both GUI and CLI. I have used command "hosted-engine --console" to start a console but error message "Cannot find character device <null>" was prompted. I also checked the status of the Hosted Engine and the its state was "EngineStarting". However, the Hosted Engine will keep restarting in the same host automatically. Please help 1. Is there any commands that I can use to start the VM inside the Hosted Engine CLI. 2. Does anyone know why my Hosted Engine cannot be started successfully and accessed via CLI. Thank you for your help. KT

2 2

Please let me know how to resolve the below error!
by d_hashima＠sagaidc.co.jp 12 Jun '25

12 Jun '25

Error Content [ ERROR ] fatal: [localhost -> 192.168.1.252]: FAILED! => {"changed": false, "msg": "Failed to download metadata for repo 'centos-ceph-pacific': Cannot prepare internal mirrorlist: Curl error (6): Couldn't resolve host name for http://mirrorlist.centos.org/?release=8-stream&arch=x86_64&repo=storage-cep… [Could not resolve host: mirrorlist.centos.org]", "rc": 1, "results": []} environment CentOS Stream release ovirt-node-ng-installer-4.5.5-2023113015.el9.iso I am building it with ovirt-node-ng-installer-4.5.5-2023113015.el9.iso and plan to install oVirt Engine there. I am building it on a virtual server.

5 10

What is the future of oVirt?
by Giacomo Failla 10 Jun '25

10 Jun '25

Hi everyone, I’m writing this post because I think many of us are asking the same questions these days. Over the past few months, we’ve seen some big changes in the open source virtualization world. Red Hat’s decision to discontinue RHV and slow down upstream oVirt development has understandably created a lot of uncertainty — especially for those of us running critical infrastructure on top of oVirt. In this changing landscape, I’m wondering: how is the oVirt project doing today? Is there still an active roadmap? Are there developers and organizations continuing to contribute? And most importantly — what can we as a community do to help ensure that oVirt remains a solid and well-maintained option in the years ahead? At the end of the day, many of us need to make decisions about how to evolve our IT environments, and having clarity on whether oVirt will continue to be a reliable choice is really important. I’d love to hear from anyone close to the project or involved in development. Any updates, insights, or thoughts would be greatly appreciated. Thanks to everyone for your work and contributions to this community. Best regards, Giacomo Failla Head of Technology — Cheope SA

2 1

Can't install the self-hosted engine without mirror or missing package errors
by andersonit＠gmail.com 06 Jun '25

06 Jun '25

I have been trying to get a working oVirt install of the self-hosted engine for the past 2 days. I've tried AlmaLinux 9, Rocky Linux 9 and CentOS Stream 9. I wanted it to be a "production ready" build and not use CentOS or the nightly oVirt repository. I am constantly getting repository errors like "unable to resolve mirrorlist.centos.org" even though I've edited the repo files and verified none of them contain that mirror. My most recent attempt with a fresh install of CentOS Stream 9 and the nightly repository (from here https://www.ovirt.org/develop/dev-process/install-nightly-snapshot.html) ends with error: [ ERROR ] fatal: [localhost -> 192.168.1.13]: FAILED! => {"changed": false, "msg": "Failed to download metadata for repo 'centos-ceph-pacific': Cannot prepare internal mirrorlist: Curl error (6): Couldn't resolve host name for http://mirrorlist.centos.org/?release=8-stream&arch=x86_64&repo=storage-cep… [Could not resolve host: mirrorlist.centos.org]", "rc": 1, "results": []} I'm beyond frustrated and wondering how anyone is able to get this installed. I've tried multiple walkthroughs and have tried following the oVirt documentation as carefully as I can, but continually get errors. Can anyone point me to a proven installation guide that does not error out? Or, is no one actually installing the self-hosted engine and there's some other way I should be doing this? I was hoping to use Rocky Linux and the standard master oVirt repo rather than CentOS and the nightly repo in order to have a more stable system with less chance of development bugs. Please help! Thanks in advance.

1 0

ovirt-hosted-engine deploy hangs
by bleuze＠enkon.com 02 Jun '25

02 Jun '25

This is on a brand new Oracle Linux ver8.10 install on a physical server. I have completed all the perquisite steps in the instructions at https://docs.oracle.com/en/virtualization/oracle-linux-virtualization-manag… I run the deployment script: hosted-engine --deploy --4 Answer all the questions and let it run. It runs for about an hour and then hangs at the step: [ INFO ] TASK [ovirt.ovirt.hosted_engine_setup : Create temporary lock file] [ INFO ] changed: [localhost -> localhost] [ INFO ] TASK [ovirt.ovirt.hosted_engine_setup : Pause execution until /tmp/ansible.yo6aij1p_he_setup_lock is removed, delete it once ready to proceed] The first time I let it run all night and it was still hung in the morning. I have tried this using the cli command above. I have also tried it using the cockput gui method, both procedures hang at the same place. I have to run /usr/sbin/ovirt-hosted-engine-cleanup after each failure before trying again I can't find any clues in the logs but perhaps someone can tell me what to look for. I am happy to post the logs here as well. I assume this is at the point where the temporary Self Hosted engine has been created and is ready to be migrated to the shared storage. The lock placed there until it is successfully moved. But the instructions say that I should be prompted for the "storage you would like to use" I don't get that promtp. It is hung somewhere before it gets to that stage. What I do know is that the temporary Self hosted engines is created because I can ssh into and login as root Next I will try again but this time use the --ansible-extra-vars option to fill in all the variables at the beginning. I don't know what else to try

1 1

Unable to access the Console via web browser.
by natchawi28＠gmail.com 26 May '25

26 May '25

I'm currently experiencing a usage issue. I'm unable to access the Console via web browser. A banner appears at the top stating: 'Something went wrong, connection is closed.' Upon further investigation, I found that the ovirt-websocket-proxy service shows the following message: handler exception: [SSL: TLSV1_ALERT_UNKNOWN_CA] tlsv1 alert unknown ca (_ssl.c:897) OS Version: Oracle 8.10 Ovirt engine: 4.4

2 1

Potential CentosStream 9 resilientstorage repo issues
by lmd＠gto.net 26 May '25

26 May '25

Good Morning! I wanted to report that I think there are issues with CentOS Stream 9 ResilientStorage repos (added by package "ovirt-release-ovirt45" and also enabled in the hosted-engine appliance) that causes anything related to DNF and this repo to fail as captured below. Digging in, it looks like centos.org is advertising a metalink.xml (at least for my Toronto, Ontario location) that is timetamped 1747062205 (Monday, May 12, 2025 3:03:25 PM GMT). Mirrors are not in good shape: * centos.org themselves are advertising a repomd.xml for ResilientStorage with timestamp 1747752818 (Tuesday, May 20, 2025 2:53:38 PM) * ca.mirrors.cicku.me are advertising an identical repomd.xml of centos.org * mirror.cpsc.ucalgary.ca is advertisting a 0-byte repomd.xml * mirrors.lug.mtu.edu is advertising an old repomd.xml of timestamp 1744645925 (Monday, April 14, 2025 3:52:05 PM) The SHA512 sum of all published repomd.xml files does not match metalink, therefore all mirrors fail. I haven't tried to exclude the repo from the HE but I wanted to post a note here in case someone else is running into this. I anticipate that Centos.org will fix this at some point, but based on dates it could be that this has been broken for ~3 days already (2025-May-20 2:53PM GMT). I am not sure where to post to raise awareness with CentOS.org or the group responsible for the ResilientStorage repo. If someone knows please let me know - or if someone on the oVirt team could notify a contact, I would be obliged if you could let them know. Thank you kindly, Dan > Errors during downloading metadata for repository 'resilientstorage': - Downloading successful, but checksum doesn't match. Calculated: 87c76b87fe36bfa60fcd8edbc52934559d8af59850efa3f0c04904e0a757d35d9d9bc0edb6114126971709afa4670d77d2de501c949a4f5e5b6144409f7efa77(sha512) Expected: 17ef9ff2c8fc04667ed55a33dc02476a59e236e7da7f19d86c586e2fbab75fa1bd7b1e68d1a1ad4657e84c2791f5f0bf78f30be53db8b5eb7fc83227c819185f(sha512) - Downloading successful, but checksum doesn't match. Calculated: 45c8d2dfc4fe245817d3f5ce5be99505a841b1d328ef14c12ba3b789277bf271a2e36cb981ce66cd0c30f652670504f40907dc05aa8bc9858d506b1989093846(sha512) Expected: 17ef9ff2c8fc04667ed55a33dc02476a59e236e7da7f19d86c586e2fbab75fa1bd7b1e68d1a1ad4657e84c2791f5f0bf78f30be53db8b5eb7fc83227c819185f(sha512) Error: Failed to download metadata for repo 'resilientstorage': Cannot download repomd.xml: Cannot download repodata/repomd.xml: All mirrors were tried [devops@l51amew9]$

2 4

Cannot create a data nfs domain.
by mobiyef772＠daxiake.com 26 May '25

26 May '25

I cannot create a data nfs domain, but I can create a iso nfs domain. When I check my nfs share on hosts: >> python3 nfs-check.py 192.168.80.10:/var/nfs/data/ovirt-images >> Current hostname: ovirt-n1.lab.io - IP addr 10.0.65.20 >> Trying to /bin/mount -t nfs 192.168.80.10:/var/nfs/data/ovirt-images... >> Executing NFS tests.. >> Removing vdsmTest file.. >> Status of tests [OK] >> Disconnecting from NFS Server.. >> Done! In web I get: Error while executing action New NFS Storage Domain: Unexpected exception, but ovirt-engine created: ovirt-images: |-130c14d7-3d83-410e-a862-2232f5327eea |---dom_md |---images in log: >>> engine.log:2025-05-26 11:46:45,515+03 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.DisconnectStorageServerVDSCommand] (default task-14) [52d4985c-0ed1-4b7f-aa67-d2d72c1847fb] START, >>> DisconnectStorageServerVDSCommand(HostName = ovirt-n1.lab.io, StorageServerConnectionManagementVDSParameters:{hostId='da7c47b1-0e7d-42be-b3da-19ea6e513e5f', storagePoolId='00000000-0000-0000-0000-000000000000', storageType='NFS', connectionList='[StorageServerConnections:{id='70506050-af3e-4061-aca2-c80d7bcb1893', connection='192.168.80.10:/var/nfs/data/ovirt-images', iqn='null', vfsType='null', mountOptions='null', nfsVersion='AUTO', nfsRetrans='null', nfsTimeo='null', iface='null', netIfaceName='null'}]', sendNetworkEventOnFailure='true'}), log id: 6d537e69 >>> engine.log:2025-05-26 11:46:47,813+03 INFO [org.ovirt.engine.core.vdsbroker.vdsbroker.DisconnectStorageServerVDSCommand] (default task-14) [52d4985c-0ed1-4b7f-aa67-d2d72c184L7fb] FINISH, DisconnectStorageServerVDSCommand, return: {70506050-af3e-4061-aca2-c80d7bcb1893=0}, log id: 6d537e69 Why?

1 0