Re: [ovirt-users] Kernel panic for live images
I've (tentatively) confirmed this as an issue with the CentOS IRC channel - that images need to be dd'd not written with tools, meaning that this portion of the wiki is irrelevant: http://www.ovirt.org/OVirt_Live#USB On 02/08/14 09:50, Jaco wrote: update: It seems that there's a bug in the implementation of UNetBootIn and/or liveusb-creator that's causing identical kernel panics to what I'm experiencing. http://bugs.centos.org/view.php?id=6865#c20214 (I'm in the process of migrating off Ubuntu/Debian/.deb for my desktop/infrastructure to RPM's - oVirt being a crucial 1st step) Initially I suspected fault RAM (since it;s the only aspect of kernel panics I can actually do something about without chucking out my whole box), but new RAM a new host makes no difference. The resolution seems to be to do it by hand - dd. Will try this report back. - J On 01/08/14 07:28, Jaco wrote: Hi folks, I get kernel panics when booting the live images: * ovirt-live-3.4.0.el6ev.iso (MD5: 01a83d96a9650dd3da19cc37edaa7b19) * ovirt-live-el6-3.4.3-1.iso (MD5: 816fd6b1294f4619e205f2efd92f0f2f) I've run Memtest86 (v2.01) it checks out OK, the current CentOS 6.5 on the host runs OK live-booting other images (either LiveUSB or PXE, CentOS, Fedora, Debian, Ubuntu other) works fine. Has anyone else gotten this? Just seeking confirmation that it's not just me before filing a bug. I believe a new images may become available after the testing this week. Will try earlier images to compare. - J ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] Kernel panic for live images
update: It seems that there's a bug in the implementation of UNetBootIn and/or liveusb-creator that's causing identical kernel panics to what I'm experiencing. http://bugs.centos.org/view.php?id=6865#c20214 (I'm in the process of migrating off Ubuntu/Debian/.deb for my desktop/infrastructure to RPM's - oVirt being a crucial 1st step) Initially I suspected fault RAM (since it;s the only aspect of kernel panics I can actually do something about without chucking out my whole box), but new RAM a new host makes no difference. The resolution seems to be to do it by hand - dd. Will try this report back. - J On 01/08/14 07:28, Jaco wrote: Hi folks, I get kernel panics when booting the live images: * ovirt-live-3.4.0.el6ev.iso (MD5: 01a83d96a9650dd3da19cc37edaa7b19) * ovirt-live-el6-3.4.3-1.iso (MD5: 816fd6b1294f4619e205f2efd92f0f2f) I've run Memtest86 (v2.01) it checks out OK, the current CentOS 6.5 on the host runs OK live-booting other images (either LiveUSB or PXE, CentOS, Fedora, Debian, Ubuntu other) works fine. Has anyone else gotten this? Just seeking confirmation that it's not just me before filing a bug. I believe a new images may become available after the testing this week. Will try earlier images to compare. - J ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] Kernel panic for live images
Hi folks, I get kernel panics when booting the live images: * ovirt-live-3.4.0.el6ev.iso (MD5: 01a83d96a9650dd3da19cc37edaa7b19) * ovirt-live-el6-3.4.3-1.iso (MD5: 816fd6b1294f4619e205f2efd92f0f2f) I've run Memtest86 (v2.01) it checks out OK, the current CentOS 6.5 on the host runs OK live-booting other images (either LiveUSB or PXE, CentOS, Fedora, Debian, Ubuntu other) works fine. Has anyone else gotten this? Just seeking confirmation that it's not just me before filing a bug. I believe a new images may become available after the testing this week. Will try earlier images to compare. - J ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] ovirt node iso?
Cheers - I'll look for it give it a go. Could I ask if an install-to-host option is available? Also that the memtest image integrity tests are available at boot. I've also noted that there are not corresponding MD5 sums of images available on sites, so it's tricky to validate/verify the integrity of images downloaded before flashing. Please advise Kind regards - Jaco On 30/07/14 01:56, Fabian Deutsch wrote: Hey Jorick, I really appreciate that you are looking for a new ISO. We were actually abotu to release a new ISO, but then we discovered an issue with SELinux. We hope to be able to release a new iso shortly. Greetings fabian ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] REQUEST: MD5 for images
Will log feature-request ticket BTW, here's the hash: $ md5sum ovirt-live-el6-3.4.3-1.iso 816fd6b1294f4619e205f2efd92f0f2f ovirt-live-el6-3.4.3-1.iso Cheers ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] Setting up all-in-one
Cheers thanks for the reply. On 29/07/14 17:52, Yedidyah Bar David wrote: Well, hopefully you'll find oVirt best :-) cheers :) These snippets are not enough to see *which* service failed to start - you should look a few lines before that. I searched for the [ERROR] bit looked around that, but I'll take another look, thanks If it's the engine itself, you should look at its logs (/var/log/ovirt-engine/*.log). If it's vdsm, its own (/var/log/vdsm), etc. Oh, gosh! Now it's coming back to me re the interconnected bits where they failed the last time 'round Did this include removing and reinstalling vdsm/libvirt? If not, you might want to try that before reinstalling the OS. Yea - I did. I noticed in the log it ref's Python 2.6, where 2.7 is much more current, making me question if it's something underlying. They are Poc/Demo only. I wouldn't say highly unstable, but prototyping wasn't a design consideration. Basically these images are very similar to what you'll get when installing a clean OS with all-in-one, with a few differences intended to automate everything, so that the user will not have to answer anything. Note that engine-setup is ran by it after boot, the engine is not pre-setup in the image. Think I'll try a live USB instance to see what the system is capable of a semi-stable setup should look like wipe redo the host if the issue persists. Thanks for the help kind word. I'll report back Cheers - J ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] REQUEST: MD5 for images
I downloaded the latest Live ISO image I could find - ovirt-live-el6-3.4.3-1.iso - but it's not working as expected. I'd like to request form the repo maintainers to add MD5sum's to the service so that I can verify the integrity of the images I'm running. Cheers - Jaco ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] Setting up all-in-one
unstable), or would it be stable enough for in-house (Docker) prototyping before deploying to a hosted/stable/production environment? Cheers - Jaco ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Starting VM gets paused
I got hit up with the same issue - I think it had something to do with latency or timeout to my NAS/SAN, despite an isolated prioritised storage network. I'm guessing the VDSM puts VM's in a paused state whenever it hits a slight panic or speed-wobble. I've, for the time-being, moved to a NFS/POSIX share the issues has not presented again. - J On 22/03/13 22:59, Nicolas Ecarnot wrote: Hi, After two months of stable usage of this 3.1 oVirt setup, here comes the first blocking issue for which I've no other mean to ask some hint. When I'm starting a VM, the start process is running fine. Being fast enough, we can ssh-connect to it, but 5 seconds later, the VM is paused. In the manager, I see that : 2013-03-22 09:42:57,435 ERROR [org.ovirt.engine.core.vdsbroker.vdsbroker.VdsBrokerObjectsBuilder] (QuartzScheduler_Worker-40) Error in parsing vm pause status. Setting value to NONE 2013-03-22 09:42:57,436 INFO [org.ovirt.engine.core.vdsbroker.VdsUpdateRunTimeInfo] (QuartzScheduler_Worker-40) VM serv-chk-adm3 3e17586d-bf8f-465b-8075-defaac90bc95 moved from PoweringUp -- Paused And on the host, I see one warning message, no error msg, and many looping repeated messages : * Warning : Thread-1968::WARNING::2013-03-22 09:19:18,536::libvirtvm::1547::vm.Vm::(_readPauseCode) vmId=`3e17586d-bf8f-465b-8075-defaac90bc95`::_readPauseCode unsupported by libvirt vm * Repeated msgs, amongst other repeated ones : Thread-1973::DEBUG::2013-03-22 09:19:20,247::libvirtvm::220::vm.Vm::(_getNetworkStats) vmId=`3e17586d-bf8f-465b-8075-defaac90bc95`::Network stats not available Thread-1973::DEBUG::2013-03-22 09:19:20,247::libvirtvm::240::vm.Vm::(_getDiskStats) vmId=`3e17586d-bf8f-465b-8075-defaac90bc95`::Disk hdc stats not available I made my homework and found some bugs that could be similar : https://bugzilla.redhat.com/show_bug.cgi?id=660598 https://bugzilla.redhat.com/show_bug.cgi?id=672208 and moreover : https://bugzilla.redhat.com/show_bug.cgi?id=695393 - I tried to restart the node's vds daemon : same behavior - I tried to reboot the node : same behavior - I tried to to restart the manager's engine : same behavior - I tried to run this VM on another node : same behavior - I tried to run another VM on the node I saw the issue : the other VM is running fine. I don't know if I have to conclude that this issue is specific to this VM, but I sounds like yes. Things to say about this VM : - it it a RH6 IIRC. It has already been successfully started, migrated, stopped and rebooted many times in the past. - it has 3 disks : one for the system and two for datas. - it has no snapshots - it has no different or complicated network setup My storage domain is a SAN, iSCSI linked, and doing good job since months. I must admit I'm a bit stuck. Last thing I haven't tried is to reboot the manager, though I'm not sure that would help. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] iSCSI Storage mirror
Hi Tal, Thanks for the reply. The process I was considering following was something like this: http://koo.fi/blog/2009/06/10/redundant-iscsi-storage-for-linux/ Write my VM images to a local disk then use an iSCSI target to maintain a copy in the event that the host goes offline, or if I otherwise need to do a migration. I have been looking into GlusterFS (another system I have little knowledge or experience of), and I believe that it's more of a network-based highly-available -resilient system than one designed for performance. I'm also a little loathe at this late stage of my build to introduce more subsystems that'll to be planned for, troubleshooted maintained - I'm not so sure my existing setup will have good support for it anyway. It is certainly something worth-while to consider, but just not at this stage (probably v.2 of my deployment). I hope this answers your questions. Kind regards - Jaco On 13/3/14 5:49 , Tal Nisan wrote: As for the oVirt point of view, oVirt does not support storage mirroring out of the box unless it used with Gluster FS As for the iSCSI mirroring, what exactly do you mean by mirroring? If you refer to replication then you should probably use Gluster for that matter, if not, can you please send me the actions you are to do manually to try and achieve this mirroring so I can understand better On 03/09/2013 11:40 PM, Jaco wrote: Hi guys, Got my system (3.2 on CentOS 6.3 - PoC lab v.0.9) working fairly OK, with a few issues that's of concern to me. I've now started using iSCSI storage, but kept running into an issue where the VM's would go into a paused state. A bit of digging in the logs show that it's because of a timeout issue talking to the iSCSI server/target, which for me raises the spectre of potential corruption, especially under load. Couldn't understand how this was possible, as I went out bought some dedicated hardware to set up a totally separate isolated storage network, but ended up simply running a cross-over UTP between the machines (process-of-elimination all that), but the issue persisted. This morning I found that one of the mirrored drives started failing, so (until I've replaced the drive discovered otherwise) I suspect that may be the possible cause of the issue. What occurred to me last night, as this thing was keeping me awake, is that this might not be the *best* course of action, and started thinking that maybe another way of doing it, especially since oVirt does some fairly low-level LVM stuff, is to rather store the VM's on a local drive, get far better IOPS than I could hope for with iSCSI over GBE, and rather set up the iSCSI to mirror the local device. That way the data is still available on the target in the event a fail-over/migration needs to take place, but that I'm reducing the risks a bit while improving overall performance. Is there a way to do this via oVirt, or would I have to do it manually by setting up storage locally set up the mirroring via iSCSI manually as an OS-level? And if so, what would I be looking for what sort of caveats would I have to keep in mind in order to make this setup suitable for use by multiple hosts in the event a (live-)migration needs to take place? (I'm pretty new at the iSCSI-thing LVM knowledge is just passable) I'd appreciate anyone's insights into this subject Kind regards - Jaco ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Cluster, hosts, everything down and totally stuck
does vdsm run ok on the host (you can check by running this in the host: vdsClient 0 getVdsCapabilities) I've been seeing this issue on my PoC box I've been tinkering on - for some reason the VDSM daemon is not starting up at boot (need to look into the cause of this), but once I manually start/restart it, the rest seems to carry on OK - J ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Cluster, hosts, everything down and totally stuck
Check if unused nics are using dhcp? Nope - I have 2 NIC's, both statically locked. 1st NIC for bridge, 2nd dedicated to storage traffic; direct UTP to my NAS. I have a suspicion that there's issue with the storage not coming up exactly as expected. I'm considering addressing the issue by making use of local storage rather than iSCSI, and then rather use LVM to mirror the local disk to the iSCSI. (^detailed in another messgase) - J ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] VNC web console
http://cloudiad.com/analysis/doc1 it looks noVNC would be easier to integrate with. Thanks for the link - did not know that. I think noVNC has a bit more traction in the industry (with good reason), since is does seem to be a fairly simple straight-forward client, whereas Guagamole is actually a more generic/general VNC ( RDP) proxy. In theory Guacamole may be extendible to the SPICE protocol, enabling a single interface for accessing instance consoles. - J ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[Users] VNC web console
I've seen this in the list archives, but not sure how much traction it's gained in the intervening period (or ). I've fiddled with, and had some pretty decent experiences with Guacamole: http://guac-dev.org/ An HTML(5?)-based VNC client for a browser - no client-side components needed :) Could fit in nicely with the project. Just putting it out there - J ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[Users] Failing on new setup
Hi guys, New to oVirt - and new to Fedora after a *very long* break @ using Debian/Ubuntu almost exclusively for the last few years. Much has changed, so please pardon my ignorance - got some catching-up to do. I'm doing PoC of a number of systems, so need to give oVirt a fair shake, but I'm tripping up before I can even get going (does not bode well). I've tried following the QuickStart, all-in-one other guides, quickly came across issues. Next I'll have to start building from source - a prospect that will by no means present me with less issues. New F18, minimal headless server. Installed via yum as documented, but install failed at the last steps, as there seems to be an issue with the scripts' implementation of firewalld. A very nice chap @ the IRC channel managed to help be out to get over that initial hurdle - had to *completely* disable firewalld select the 'None' option for the setup. (seems like a really bad idea to start off like that - disabling security - but this is a lab PoC, so OK, will let it slide for now) I've also disabled SELinux IPtables via systemctl - so that we can eliminate that as an issue. Script(s) still don't complete without issue, but at lease the engine seems to come up I can log in. After logging in, I try to (re)connect to the local hypervisor via (root@) $FQDN, localhost 127.0.0.1, but it fails to come up (why root? OK). It adds the host, but there seems to be connectivity issues. I'm seeing this in the GUI events: Host localhost installation failed. Command returned failure code 1 during SSH session 'root@localhost'. I'm seeing this in `tail -f /var/log/ovirt-engine/engine.log` http://pastebin.com/nVyNGSsf Relevant line being: ERROR [org.ovirt.engine.core.utils.ssh.SSHDialog] (pool-3-thread-31) SSH error running command root@localhost:'umask 0077; MYTMP=$(mktemp -t ovirt-XX); trap chmod -R u+rwX \${MYTMP}\ /dev/null 21; rm -fr \${MYTMP}\ /dev/null 21 0; rm -fr ${MYTMP} mkdir ${MYTMP} tar -C ${MYTMP} -x ${MYTMP}/setup DIALOG/dialect=str:machine DIALOG/customization=bool:True': java.io.IOException: Command returned failure code 1 during SSH session 'root@localhost' This is a *fresh* box - I've reinstalled it a few times from scratch now - I've re-installed oVirt through the manners documented, but I'm not getting anywhere fast. In this, google is not my friend, as results are sparse not very helpful - i.e. if someone's found resolution, it has not been posted. Either there's a bug in the docco's or the code/scripts. Please help - I'd *really* like to give this a go Kind regards -Jaco ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users