- Users - oVirt List Archives

Connection issues when using gluster + infiniband + RDMA
by housemouse＠squeakz.net 13 Aug '18

13 Aug '18

I am trying to setup gluster using Infiniband and mounting it in RDMA mode. I have created a network hook so that it configures the interfaces as MTU=65520 and CONNECTED_MODE=yes. I have a second server doing NFS with RDMA over Infiniband with some VMs on it. When I try and transfer files to the gluster storage it is taking a while and I am seeing the message “VDSM {hostname} command GetGlusterVolumeHealInfoVDS failed: Message timeout which can be caused by communication issues” This is usually followed by “Host {hostname} is not responding. It will stay in Connecting state for a grace period of 60 seconds and after that an attempt to fence the host will be issued.” I just installed the Infiniband hardware. The switch is a Qlogic 12200-18 with QLE7340 single port Infiniband cards in each of the 3 servers. The error message varies on which of the 3 nodes it comes from. Each of the 3 servers is running the opensm service.

3 2

Passing through Quadro cards to guest
by Wesley Stewart 13 Aug '18

13 Aug '18

Has anyone experimented with this? I have had really good success with AMD cards, but consumer end graphics cards fail to load due to NVIDIA blocking them. I happened to have an NVIDIA Quadro P2000 lying around and I thought I would give it a shot. I really like these cards as they only require the 75W provided PCI power. I can pass the Quadro through just fine, however after a couple minutes of playing a video, the audio gets really "buzzy" and the video drops to about 3-5 FPS. This is while doing pretty minimal stuff. Currently testing with 720/1080 youtube videos. Card is pulling about 6-10W total. And otherwise working fine on my Windows 10 guest. I have also tried enabling/disabling MSI interrupts which the card supports without making a difference. I am running a single host setup: X10SDV-6c+ (Xeon D - 1528) running oVirt 4.2.3

1 0

Connection issues when using gluster + infiniband + RDMA
by Matthew Hoberg 13 Aug '18

13 Aug '18

I am trying to setup gluster using Infiniband and mounting it in RDMA mode. I have created a network hook so that it configures the interfaces as MTU=65520 and CONNECTED_MODE=yes. I have a second server doing NFS with RDMA over Infiniband with some VMs on it. When I try and transfer files to the gluster storage it is taking a while and I am seeing the message “VDSM {hostname} command GetGlusterVolumeHealInfoVDS failed: Message timeout which can be caused by communication issues” This is usually followed by “Host {hostname} is not responding. It will stay in Connecting state for a grace period of 60 seconds and after that an attempt to fence the host will be issued.” I just installed the Infiniband hardware. The switch is a Qlogic 12200-18 with QLE7340 single port Infiniband cards in each of the 3 servers. The error message varies on which of the 3 nodes it comes from. Each of the 3 servers is running the opensm service. Thanks, Matt

1 0

Isolated POC
by Martin Schutte 13 Aug '18

13 Aug '18

Good day to all, I am in the process of setting up an oVirt proof of concept for a client. There are some strict requirements for this POC. Firstly the entire POC must be isolated from the network, which means no online repos. Secondly there is a time constraint for the completion of the deployment. All to be done on bare-metal systems. I have installed oVirt numerous times on other systems, but then I was always able to have online repos. So in short I am asking for some pointers on which repos needs to be synced offline for this to work properly. I can always go and play around to figure out which ones are needed, but I do not have the luxury of time. Any advice or tips will be appreciated. Regards, Martin.

5 8

oVirt node 4.2.x - custom multipath.conf and external storage availability
by Roberto Nunin 13 Aug '18

13 Aug '18

HI all How can I provide to my oVirt hosts a custom multipath.conf, considering that we are using shared/mirrored FC storage from HPE 3PAR with high availability of provided LUNs, using peer-persistance feature ? To configure correctly this require a custom multipath.conf is required. How I can avoid it will be overwritten during upgrades ? Second topic: Is ALUA feature supported also in oVirt nodes, like for standard RHEL 6/7 and/or CentOS 6/7 ? I was digging for a while into user's digest, but not able to find focused informations. This is the feature needed by peer-persistance, allowing two different volumes to be masked toward a given host with the same WWN and LUN number. Thanks in advance -- Roberto

2 2

ovirt engine service needs restart after network connectivity is lost and reestablished
by g.vasilopoulos＠uoc.gr 13 Aug '18

13 Aug '18

today our network administration did some upgrades on the networking equipment, so the engine vlan went down for a while. Afterwards when it came back up 3 hosts was found as non responding, I couldn't see anything suspicious on the hosts, the problem "fixed" itself when I restarted the ovirt-engine service on the engine vm. Another "strange" side effect is that gluster was in need of healing constantly afterwards it did heal and desynch again almost immediately, whitch was "fixed" by shutting down and starting the vm's who were up at the time of the network outage.. I've seen this behaviour once more regarding the engine service, is this normal behaviour (e.g. to make the admin check if everything is ok) or is there something wrong ?

2 1

cannot activate host
by Jiří Sléžka 11 Aug '18

11 Aug '18

Hi, I have one interesting issue. I was asked by colleague to repair his cluster which is part of our oVirt installation. He tried to upgrade his host to gain 4.2 compatibility level but it fails in some cases. Never mind, I checked it, repaired, installed upgrades and activate his hosts with one exception. One of hosts was stucked in finishing two tasks. One was changing network configuration, the second was activating of this host. I have to manually clear this tasks in database by SELECT * FROM job ORDER BY start_time DESC; SELECT DeleteJob('...job_id...'); SELECT DeleteJob('...job_id...'); Now this tasks are gone but still cannot activate this host. Error while executing action: blade01: Cannot activate Host. Related operation is currently in progress. Please try again later. Host is updatet (latest 4.2 repo), rebooted, looks fine... In tasks in UI are no tasks found (the same in db)... What can I do now? Cheers, Jiri

5 11

Digital Marketing Proposal
by Manya Avasti 11 Aug '18

11 Aug '18

Dear Sir/Madam, Do you know social media marketing has a 100% higher lead-to-close rate than outbound marketing? Or that 84% of B2B marketers use social media in some form? Social Media marketing is the best platform for generate referral traffic & conversion. We have an expert social media marketing team and we promote your online products and brand on social media platforms. What we do on social media platforms: *Facebook* · Facebook a/c Page & Fan Page Set-up · Theme Related Cover Photo · Facebook Groups posting · Facebook Likes *Twitter* · Twitter a/c Set-up · Tweets on profile · Background Photo upload according to theme/business · Twitter Followers improvement *Linkedin* · Linkedin a/c Set-up · Post content on profile or company page. · Company page creation · Theme Related Cover Photo · Increase the Connections *Google Plus* · Google + a/c Set-up · Posts content on profiles · Theme Related Cover Photo implementation · Increase Followers and Circles *Pinterest* · Pinterest a/c Set-up · Pinterest Boards creation · Pinterest Updates/Month · Followers increment If you are interested in our service then kindly revert me back. Thanks & Regards Manya

1 0

Reading an ISO image to identify the Image version and type
by Hari Prasanth Loganathan 10 Aug '18

10 Aug '18

Hi Team, I have few ISO images, Is there a utility in oVirt which tells us which image version it is (like windows / linux version)? 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.

3 6

Re: Live Migration via NFS
by Douglas Duckworth 10 Aug '18

10 Aug '18

Hi Nir Thanks for clarifying We can still use NFS as export domain which can be used to move VM once they're shutdown? Can GPFS be used to provide shared block storage domain? Thanks, Douglas Duckworth, MSc, LFCS HPC System Administrator Scientific Computing Unit Weill Cornell Medicine E: doug(a)med.cornell.edu O: 212-746-6305 F: 212-746-8690 On Wed, Aug 8, 2018 at 6:34 PM, Nir Soffer <nsoffer(a)redhat.com> wrote: > On Wed, Aug 8, 2018 at 4:44 AM Douglas Duckworth <dod2014(a)med.cornell.edu> > wrote: > >> Can NFS facilitate live migration in cluster of two hosts each which have >> NFS mounted? >> >> The hosts would have local attached storage which would be the original >> location of VMs. >> > > We don't support migrating vm migration from local to shared storage, only > if the disk is > on shared storage. > > Nir >

3 3