- Users - oVirt List Archives

high load on hosts
by Jayme 24 Oct '18

24 Oct '18

I'm wondering how I can best limit the ability of VMs to overrun the load on hosts. I have a fairly stock 4.2 HCI setup with three well spec'ed servers, 10Gbe/SSDs, plenty of ram and CPU with only a hand full of light use VMs. I notice when the occasional demanding job is run on a VM I'm seeing load average on host node shoot up in to the 20-30s, how can a single "medium" vm cause host load to rise so high?

4 6

Choosing optimal vdo Write policy on LSI Raid device that always report write-trough to the OS
by guillaume.pavese＠interact-iv.com 24 Oct '18

24 Oct '18

Hi, I am evaluating vdo for an hyperconverged ovirt cluster. The volume is on RAID device supported by a LSI (Dell Perc H710) controller with Write Back cache enabled : Current Cache Policy: WriteBack, ReadAdaptive, Cached, Write Cache OK However these LSI controllers (all PERC series on Dell servers), as per documentation, always report cache policy as writethrough to the OS, no matter if write back is effectively enabled or not : cat /sys/block/sdb/device/scsi_disk/0\:2\:1\:0/cache_type write through vdo by default choose its write policy (sync or async) accoring to the cache mode reported by the disk. So with cache policy auto, in that case it is set to "sync" : write policy: sync So my question : In this configuration, write cache enabled on the raid controller but reported as write trhough to the os, should I manually set vdo cache policy to async for optimal performance? Thanks for your help

1 0

Re: kvm import failed with error: Operation not supported: peeking is supported only for RAW disks
by Karli Sjöberg 23 Oct '18

23 Oct '18

1 0

Ovirt 4.2 rpm glusterfs-gnfs w/ gluster 4.1
by TomK 23 Oct '18

23 Oct '18

Hey All, Is there a newer package of glusterfs-gnfs available for GlusterFS 4.1? After upgrading to GlusterFS 4.1, all the hosts are now disconnected from the Ovirt engine. I'm on CentOS 7.5: ---> Package kmod-kvdo.x86_64 0:6.1.0.181-17.el7_5 will be installed ---> Package mokutil.x86_64 0:12-2.el7 will be installed ---> Package python-imgbased.noarch 0:1.0.24-1.el7 will be installed ---> Package vdo.x86_64 0:6.1.0.168-18 will be installed --> Finished Dependency Resolution Error: Package: glusterfs-gnfs-3.12.15-1.el7.x86_64 (ovirt-4.2-centos-gluster312) Requires: glusterfs-client-xlators(x86-64) = 3.12.15-1.el7 Installed: glusterfs-client-xlators-4.1.5-1.el7.x86_64 (@centos-gluster41) glusterfs-client-xlators(x86-64) = 4.1.5-1.el7 Available: glusterfs-client-xlators-3.8.4-53.el7.centos.x86_64 (base) glusterfs-client-xlators(x86-64) = 3.8.4-53.el7.centos Available: glusterfs-client-xlators-3.8.4-54.15.el7.centos.x86_64 (updates) glusterfs-client-xlators(x86-64) = 3.8.4-54.15.el7.centos -- Cheers, Tom K. ------------------------------------------------------------------------------------- Living on earth is expensive, but it includes a free trip around the sun. [root@mdskvm-p02 yum.repos.d]# yum info glusterfs-gnfs Loaded plugins: enabled_repos_upload, fastestmirror, package_upload, product-id, search-disabled-repos, subscription-manager, vdsmupgrade This system is not registered with an entitlement server. You can use subscription-manager to register. Loading mirror speeds from cached hostfile * base: mirror.csclub.uwaterloo.ca * epel: mirror.csclub.uwaterloo.ca * extras: mirror.csclub.uwaterloo.ca * ovirt-4.2: mirrors.rit.edu * ovirt-4.2-epel: mirror.csclub.uwaterloo.ca * updates: mirror.csclub.uwaterloo.ca Available Packages Name : glusterfs-gnfs Arch : x86_64 Version : 3.12.15 Release : 1.el7 Size : 166 k Repo : ovirt-4.2-centos-gluster312/x86_64 Summary : GlusterFS gNFS server URL : http://gluster.readthedocs.io/en/latest/ License : GPLv2 or LGPLv3+ Description : GlusterFS is a distributed file-system capable of scaling to several : petabytes. It aggregates various storage bricks over Infiniband RDMA : or TCP/IP interconnect into one large parallel network file : system. GlusterFS is one of the most sophisticated file systems in : terms of features and extensibility. It borrows a powerful concept : called Translators from GNU Hurd kernel. Much of the code in GlusterFS : is in user space and easily manageable. : : This package provides the glusterfs legacy gNFS server xlator

2 2

Import VM from VMware OVA
by Hari Prasanth Loganathan 23 Oct '18

23 Oct '18

Hi Team, 1) I exported a ovf file from VMware 2) I converted the ovf file to ova using ovftool successfully. 3) I imported the ova using the VM Import -> VMware appliance OVA -> provided the path of the ova (copied in ovirt location) 4) VM is successfully imported into the ovirt and I see the disk is also successfully imported BUT the disk is stored as image disk, What is the possible way to import the VM (as mentioned above) and store the disk as Cinder disk (currently always it is storing as image disk type) Any help is much appreciated. Thanks, Hari -- DISCLAIMER - *MSysTechnologies LLC* This email message, contents and its attachments may contain confidential, proprietary or legally privileged information and is intended solely for the use of the individual or entity to whom it is actually intended. If you have erroneously received this message, please permanently delete it immediately and notify the sender. If you are not the intended recipient of the email message,you are notified strictly not to disseminate,distribute or copy this e-mail.E-mail transmission cannot be guaranteed to be secure or error-free as Information could be intercepted, corrupted, lost, destroyed, incomplete or contain viruses and MSysTechnologies LLC accepts no liability for the contents and integrity of this mail or for any damage caused by the limitations of the e-mail transmission.

1 0

Re: cloud-init behaviour after reboot and growpart [update : BUG]
by Karli Sjöberg 23 Oct '18

23 Oct '18

1 0

cloud-init behaviour after reboot and growpart
by Nathanaël Blanchet 23 Oct '18

23 Oct '18

Hi all, After having successfully run cloud-init with the "initial run" box checked, I meet an issue after rebooting the el7 vm. At the boot time, cloud-init waits for a data source that doesn't exist anymore and I have to wait serveral timeouts like this one: open 'http://169.254.169.254/openstack' with {'url': 'http://169.254.169.254/openstack', 'headers': {'User-Agent': 'Cloud-Init/0.7.9'}, 'allow_redirects': True, 'method': 'GET', 'timeout': 10.0} configuration As a workaround, I must include in the initial custom script "touch /etc/cloud/cloud-init.disabled" so as to the vm to boot properly. Now I'd like cloud-init not to be fully disabled because I need it resizes the filesystem after the growpart step. Can anyone help me with this? -- Nathanaël Blanchet Supervision réseau Pôle Infrastrutures Informatiques 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

2 3

OpenStack Block Storage Provider without authentication doesn't work
by Matthias Leopold 23 Oct '18

23 Oct '18

Hi, we're successfully using Cinder as a Block Storage Provider for oVirt with customized Cinder installation in a CentOS host according to OpenStack docs. Now i wanted to try out running Cinder in Docker and followed the instructions from https://thenewstack.io/deploying-cinder-stand-alone-storage-service/ (customized for use with Ceph RBD). This works to the point where i can setup an external provider and consequently a storage domain in oVirt. I set up the provider without authentication and testing this (by pressing "Test" button) works. When i want to create disks i realize that oVirt doesn't recognize the "Volume Type" definitions. In engine.log i see messages like 2018-10-03 12:04:46,990+02 ERROR [org.ovirt.engine.core.bll.storage.disk.cinder.GetCinderVolumeTypesByStorageDomainIdQuery] (default task-72) [06d4fbf7-0b3c-46b3-8166-148ee7f67a4c] Query 'GetCinderVolumeTypesByStorageDomainIdQuery' failed: null 2018-10-03 12:04:46,990+02 ERROR [org.ovirt.engine.core.bll.storage.disk.cinder.GetCinderVolumeTypesByStorageDomainIdQuery] (default task-72) [06d4fbf7-0b3c-46b3-8166-148ee7f67a4c] Exception: java.lang.NullPointerException I very much suspect that not using authentication (because of "auth_strategy = noauth" in cinder.conf) is the culprit. Listing types with cinder CLI from the engine host works (when using the appropriate environment). Can some of the RH devs confirm this behaviour? Is this an engine bug? thanks Matthias

2 4

Tutorial for beginners: oVirt - moving away from local storage setup to NFS/GlusterFS
by Andrei Verovski 23 Oct '18

23 Oct '18

Hi, I made short video tutorial (slide show with screenshots and instruction) for beginners: oVirt - moving away from local storage setup to shared/networked one (main currently uses NFS loopback). https://www.youtube.com/watch?v=2ebhZ5U6x_Y Hopefully someone may avoid common mistakes, and if not, remedy situation with minimal time and effort. Feedback and comments are welcome. with best regards Andrei

1 0

re-enabling networkmanager
by fsoyer 23 Oct '18

23 Oct '18

Hi, I have installed a 4.2 cluster on CentOS 7 nodes but I have follow an (old) procedure of mine done with 4.0 : so, I have disabled Network Manager before installing oVirt. The networks created and validated in the engine UI are : ovirmgmt on bond0 (2 slaves) failover mode storagemanager on bond1 (2 slaves), jumbo frames, aggregation mode, serving Gluster. Today, I installed Cockpit on the node to have the nodes consoles. But it say that it cannot manage the network without NM. So my question is : is there any risk to re-enabled NM on the nodes ? Can it broke anything done by the UI ? -- Cordialement, Frank Soyer Mob. 06 72 28 38 53 - Fix. 05 49 50 52 34 Systea IG Administration systèmes, réseaux et bases de données www.systea.net Membre du réseau Les Professionnels du Numérique KoGite Hébergement de proximité www.kogite.fr

4 6