Re: [ovirt-users] [RFI] oVirt 3.6 Planning
- Original Message - From: Sven Kieske s.kie...@mittwald.de To: users@ovirt.org Sent: Friday, September 12, 2014 4:53:46 PM Subject: Re: [ovirt-users] [RFI] oVirt 3.6 Planning +1! I really would like to see a fast working way to create, manage and use ceph in ovirt. I don't know if implementation through openstack component is enough, because you will need more components then (keystone for auth e.g.). Sure you will need more components, but we don't want to duplicate the effort, implementing ceph support directly in ovirt. Using Cinder we can reuse the existing ceph support, and gain also support for other Cinder drivers for free. On the host side, we will have of course ceph specific support, so you will be able to create ceph volumes directly, and use vdsm api to start vms using these volumes, if you want more direct support. I think this is the general long term approach; integrate openstack components instead of re-implementing the wheel. On 12/09/14 15:34, Cédric Buot de l'Epine wrote: Dunno if CephFS efforts will be sustainable against Glusterfs... I'm not sure if a posix fs is the best provider for blocks. I would be pleased to have a direct form to create a Ceph pool for a datacenter, then provide librbd path for the guests (or only their size). -- Mit freundlichen Grüßen / Regards Sven Kieske Systemadministrator Mittwald CM Service GmbH Co. KG Königsberger Straße 6 32339 Espelkamp T: +49-5772-293-100 F: +49-5772-293-333 https://www.mittwald.de Geschäftsführer: Robert Meyer St.Nr.: 331/5721/1033, USt-IdNr.: DE814773217, HRA 6640, AG Bad Oeynhausen Komplementärin: Robert Meyer Verwaltungs GmbH, HRB 13260, AG Bad Oeynhausen ___ 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: [ovirt-users] [RFI] oVirt 3.6 Planning
When VMs goes to unknown status, all control options are grayed, and the only way to fix this is to update the Database manually, it will be nice to be able to force the engine to poweroff the VM from admin portal. Thanks, Mohyedeen. On Sat, Sep 13, 2014 at 12:27 PM, Nir Soffer nsof...@redhat.com wrote: - Original Message - From: Sven Kieske s.kie...@mittwald.de To: users@ovirt.org Sent: Friday, September 12, 2014 4:53:46 PM Subject: Re: [ovirt-users] [RFI] oVirt 3.6 Planning +1! I really would like to see a fast working way to create, manage and use ceph in ovirt. I don't know if implementation through openstack component is enough, because you will need more components then (keystone for auth e.g.). Sure you will need more components, but we don't want to duplicate the effort, implementing ceph support directly in ovirt. Using Cinder we can reuse the existing ceph support, and gain also support for other Cinder drivers for free. On the host side, we will have of course ceph specific support, so you will be able to create ceph volumes directly, and use vdsm api to start vms using these volumes, if you want more direct support. I think this is the general long term approach; integrate openstack components instead of re-implementing the wheel. On 12/09/14 15:34, Cédric Buot de l'Epine wrote: Dunno if CephFS efforts will be sustainable against Glusterfs... I'm not sure if a posix fs is the best provider for blocks. I would be pleased to have a direct form to create a Ceph pool for a datacenter, then provide librbd path for the guests (or only their size). -- Mit freundlichen Grüßen / Regards Sven Kieske Systemadministrator Mittwald CM Service GmbH Co. KG Königsberger Straße 6 32339 Espelkamp T: +49-5772-293-100 F: +49-5772-293-333 https://www.mittwald.de Geschäftsführer: Robert Meyer St.Nr.: 331/5721/1033, USt-IdNr.: DE814773217, HRA 6640, AG Bad Oeynhausen Komplementärin: Robert Meyer Verwaltungs GmbH, HRB 13260, AG Bad Oeynhausen ___ 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 ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] [RFI] oVirt 3.6 Planning
One more thing,, Being able to attach USB dongles from the admin portal... Thanks, Mohyedeen. On Sat, Sep 13, 2014 at 3:32 PM, Mohyedeen Nazzal mohyedeen.naz...@gmail.com wrote: When VMs goes to unknown status, all control options are grayed, and the only way to fix this is to update the Database manually, it will be nice to be able to force the engine to poweroff the VM from admin portal. Thanks, Mohyedeen. On Sat, Sep 13, 2014 at 12:27 PM, Nir Soffer nsof...@redhat.com wrote: - Original Message - From: Sven Kieske s.kie...@mittwald.de To: users@ovirt.org Sent: Friday, September 12, 2014 4:53:46 PM Subject: Re: [ovirt-users] [RFI] oVirt 3.6 Planning +1! I really would like to see a fast working way to create, manage and use ceph in ovirt. I don't know if implementation through openstack component is enough, because you will need more components then (keystone for auth e.g.). Sure you will need more components, but we don't want to duplicate the effort, implementing ceph support directly in ovirt. Using Cinder we can reuse the existing ceph support, and gain also support for other Cinder drivers for free. On the host side, we will have of course ceph specific support, so you will be able to create ceph volumes directly, and use vdsm api to start vms using these volumes, if you want more direct support. I think this is the general long term approach; integrate openstack components instead of re-implementing the wheel. On 12/09/14 15:34, Cédric Buot de l'Epine wrote: Dunno if CephFS efforts will be sustainable against Glusterfs... I'm not sure if a posix fs is the best provider for blocks. I would be pleased to have a direct form to create a Ceph pool for a datacenter, then provide librbd path for the guests (or only their size). -- Mit freundlichen Grüßen / Regards Sven Kieske Systemadministrator Mittwald CM Service GmbH Co. KG Königsberger Straße 6 32339 Espelkamp T: +49-5772-293-100 F: +49-5772-293-333 https://www.mittwald.de Geschäftsführer: Robert Meyer St.Nr.: 331/5721/1033, USt-IdNr.: DE814773217, HRA 6640, AG Bad Oeynhausen Komplementärin: Robert Meyer Verwaltungs GmbH, HRB 13260, AG Bad Oeynhausen ___ 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 ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] [RFI] oVirt 3.6 Planning
Le 12/09/2014 14:22, Itamar Heim a écrit : With oVirt 3.5 nearing GA, time to ask for what do you want to see in oVirt 3.6? Thanks, Itamar ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users For God's sake, please, prevent secondary storage domains (anything other than the master, especially iso and export) to completely block the whole thing when being unavailable. To many users suffered from this in the last two years I participate at this mailing list (including me). -- Nicolas Ecarnot ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] : password authentication failed for user engine
[image: logo] Navteca Ramon Ramirez-Linan C: 301.789.9139 O: 202.505.1553 *www.navteca.com http://www.navteca.com/* 6301 Ivy Lane, Greenbelt, MD 20770 Suite 700 [image: script] [image: https://s3.amazonaws.com/www.navteca.com/img/certified.png] On Sat, Sep 13, 2014 at 10:32 AM, Ramon Ramirez-Linan rli...@navteca.com wrote: Hello, New to the list and relatively new to oVirt I am trying to deploy the all in one for ovirt3.4. But the engine-setup script is stuck in the process of creating the db schema reating PostgreSQL 'engine' database [ INFO ] Configuring PostgreSQL [ INFO ] Creating Engine database schema [ ERROR ] Failed to execute stage 'Misc configuration': Command '/usr/share/ovirt-engine/dbscripts/create_schema.sh' failed to execute [ INFO ] Yum Performing yum transaction rollback [ INFO ] Rolling back database schema - OperationalError: FATAL: password authentication failed for user engine .FATAL: password authentication failed for user engine Any suggestions would be greatly appreciate Thanks Rezuma ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] [RFI] oVirt 3.6 Planning
- Original Message - From: Nicolas Ecarnot nico...@ecarnot.net To: users@ovirt.org Sent: Saturday, September 13, 2014 4:53:03 PM Subject: Re: [ovirt-users] [RFI] oVirt 3.6 Planning Le 12/09/2014 14:22, Itamar Heim a écrit : With oVirt 3.5 nearing GA, time to ask for what do you want to see in oVirt 3.6? Thanks, Itamar ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users For God's sake, please, prevent secondary storage domains (anything other than the master, especially iso and export) to completely block the whole thing when being unavailable. To many users suffered from this in the last two years I participate at this mailing list (including me). I think this is true for ISO and export domains for some time. It will not work for data domains - if you have vms with disks on what you call secondary data domain, how would you migrate these vms to a host that cannot see the secondary domain? In 3.6, there will be no master domain, so any data domain will be important as any other data domain. Maybe what do you like is to have more control on which domains are critical, and which are not. A domain which you mark as non-critical, or secondary, will not cause the host to become non-operational when the host cannot see this domain. So you would not be able to migrate some vms to a host that cannot see the secondary domain, but since *you* marked it as secondary, it is not a problem for you. What do you think? Nir ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] [ovirt-devel] feature review - ReportGuestDisksLogicalDeviceName
On Wed, Sep 03, 2014 at 08:50:12AM +0200, Michal Skrivanek wrote: On Sep 2, 2014, at 21:03 , Nir Soffer nsof...@redhat.com wrote: - Original Message - From: Michal Skrivanek mskri...@redhat.com To: Liron Aravot lara...@redhat.com, Dan Kenigsberg dan...@redhat.com, Federico Simoncelli fsimo...@redhat.com, Vinzenz Feenstra vfeen...@redhat.com Cc: users@ovirt.org, de...@ovirt.org Sent: Tuesday, September 2, 2014 3:29:57 PM Subject: Re: [ovirt-devel] feature review - ReportGuestDisksLogicalDeviceName On Sep 2, 2014, at 13:11 , Liron Aravot lara...@redhat.com wrote: - Original Message - From: Federico Simoncelli fsimo...@redhat.com To: de...@ovirt.org Cc: Liron Aravot lara...@redhat.com, users@ovirt.org, smizr...@redhat.com, Michal Skrivanek mskri...@redhat.com, Vinzenz Feenstra vfeen...@redhat.com, Allon Mureinik amure...@redhat.com, Dan Kenigsberg dan...@redhat.com Sent: Tuesday, September 2, 2014 12:50:28 PM Subject: Re: feature review - ReportGuestDisksLogicalDeviceName - Original Message - From: Dan Kenigsberg dan...@redhat.com To: Liron Aravot lara...@redhat.com Cc: users@ovirt.org, de...@ovirt.org, smizr...@redhat.com, fsimo...@redhat.com, Michal Skrivanek mskri...@redhat.com, Vinzenz Feenstra vfeen...@redhat.com, Allon Mureinik amure...@redhat.com Sent: Monday, September 1, 2014 11:23:45 PM Subject: Re: feature review - ReportGuestDisksLogicalDeviceName On Sun, Aug 31, 2014 at 07:20:04AM -0400, Liron Aravot wrote: Feel free to review the the following feature. http://www.ovirt.org/Features/ReportGuestDisksLogicalDeviceName Thanks for posting this feature page. Two things worry me about this feature. The first is timing. It is not reasonable to suggest an API change, and expect it to get to ovirt-3.5.0. We are two late anyway. The other one is the suggested API. You suggest placing volatile and optional infomation in getVMList. It won't be the first time that we have it (guestIPs, guestFQDN, clientIP, and displayIP are there) but it's foreign to the notion of conf reported by getVMList() - the set of parameters needed to recreate the VM. The fact is that today we return guest information in list(Full=true), We decide on it's notion and it seems like we already made our minds when guest info was added there :) . I don't see any harm in returning the disk mapping there and if we'll want to extract the guest info out, we can extract all of it in later version (4?) without need for BC. Having the information spread between different verbs is no better imo. At first sight this seems something belonging to getVmStats (which is reporting already other guest agent information). Fede, I've mentioned in the wiki, getVmStats is called by the engine every few seconds and therefore that info wasn't added there but to list() which is called only when the hash is changed. If everyone is in for that simple solution i'm fine with that, but Michal/Vincenz preferred it that way. yes, that was the main reason me and Vinzenz suggested to use list(). 15s is a reasonable compromise, IMHO. And since it's also reported by guest agent in a similar manner (and actually via the same vdsm-ga API call) as other guest information I think it should sit alongside guestIPs, FQDN, etc… Maybe not the best place, but I would leave that for a bigger discussion if/when we want to refactor reporting of the guest agent information Given that we are late, adding disk mapping where other guest info is in a backward compatible way looks reasonable. Did you consider adding a new verb for getting guest information? This verb can be called once after starting/recovering a vm, and then only when guest info hash changes (like the xml hash). yes, was considered but turned down. Main reason is an additional overhead and pollution of the API for a tiny little item which is in the same group of guest agent reported items. It doesn't make sense to introduce an inconsistency just this one item. Refactoring of the frequency of what we report is indeed long overdue, but we shouldn't start here…(the first and most offending is still the application list) Ok, lacking another alternative, let's dump the maps to list(). ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] : password authentication failed for user engine
If this is an attempt to install over a previous installation, please try to remove /etc/ovirt-engine/engine.conf.d/*.conf, and setup again. - Original Message - From: Ramon Ramirez-Linan rli...@navteca.com To: users@ovirt.org Sent: Saturday, September 13, 2014 5:51:23 PM Subject: Re: [ovirt-users] : password authentication failed for user engine Navteca Ramon Ramirez-Linan C: 301.789.9139 O: 202.505.1553 www.navteca.com 6301 Ivy Lane, Greenbelt, MD 20770 Suite 700 On Sat, Sep 13, 2014 at 10:32 AM, Ramon Ramirez-Linan rli...@navteca.com wrote: Hello, New to the list and relatively new to oVirt I am trying to deploy the all in one for ovirt3.4. But the engine-setup script is stuck in the process of creating the db schema reating PostgreSQL 'engine' database [ INFO ] Configuring PostgreSQL [ INFO ] Creating Engine database schema [ ERROR ] Failed to execute stage 'Misc configuration': Command '/usr/share/ovirt-engine/dbscripts/create_schema.sh' failed to execute [ INFO ] Yum Performing yum transaction rollback [ INFO ] Rolling back database schema * OperationalError: FATAL: password authentication failed for user engine .FATAL: password authentication failed for user engine Any suggestions would be greatly appreciate Thanks Rezuma ___ 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: [ovirt-users] [RFI] oVirt 3.6 Planning
Le 13/09/2014 17:34, Nir Soffer a écrit : For God's sake, please, prevent secondary storage domains (anything other than the master, especially iso and export) to completely block the whole thing when being unavailable. To many users suffered from this in the last two years I participate at this mailing list (including me). I think this is true for ISO and export domains for some time. In 3.4 at least, this is still true and is a problem. At least, if we can get rid of this and just mark these iso and export domains as unusable, and still do every other non-related operations, this would be very useful. It will not work for data domains - if you have vms with disks on what you call secondary data domain, how would you migrate these vms to a host that cannot see the secondary domain? In 3.6, there will be no master domain, so any data domain will be important as any other data domain. Maybe what do you like is to have more control on which domains are critical, and which are not. A domain which you mark as non-critical, or secondary, will not cause the host to become non-operational when the host cannot see this domain. So you would not be able to migrate some vms to a host that cannot see the secondary domain, but since *you* marked it as secondary, it is not a problem for you. What do you think? This proposition is perfect. I don't know whether being able to mark some storage domain as more precious than other will be useful, when their absence won't be a blocker anymore? Nir -- Nicolas Ecarnot ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] [RFI] oVirt 3.6 Planning
- Original Message - From: Nicolas Ecarnot nico...@ecarnot.net To: Nir Soffer nsof...@redhat.com Cc: users@ovirt.org, Federico Simoncelli fsimo...@redhat.com Sent: Saturday, September 13, 2014 9:24:55 PM Subject: Re: [ovirt-users] [RFI] oVirt 3.6 Planning Le 13/09/2014 17:34, Nir Soffer a écrit : For God's sake, please, prevent secondary storage domains (anything other than the master, especially iso and export) to completely block the whole thing when being unavailable. To many users suffered from this in the last two years I participate at this mailing list (including me). I think this is true for ISO and export domains for some time. In 3.4 at least, this is still true and is a problem. At least, if we can get rid of this and just mark these iso and export domains as unusable, and still do every other non-related operations, this would be very useful. It will not work for data domains - if you have vms with disks on what you call secondary data domain, how would you migrate these vms to a host that cannot see the secondary domain? In 3.6, there will be no master domain, so any data domain will be important as any other data domain. Maybe what do you like is to have more control on which domains are critical, and which are not. A domain which you mark as non-critical, or secondary, will not cause the host to become non-operational when the host cannot see this domain. So you would not be able to migrate some vms to a host that cannot see the secondary domain, but since *you* marked it as secondary, it is not a problem for you. What do you think? This proposition is perfect. I don't know whether being able to mark some storage domain as more precious than other will be useful, when their absence won't be a blocker anymore? The idea is to let you define which storage domains are required and which are not, lets call them optional. If a required storage domain cannot be seen, a host be become non-operational. If an optional storage domain cannot be seen, you will get a warning but the host will function normally. If you try to migrate a vm to a host which cannot see the storage domain used by that vm, the operation will fail. When a required storage domain is down, you would be able to change it to optional, and continue to work with the other domains in degraded mode. Some vms will not able to run, but other vms that do not depend on problem domain will not be affected. I hope that I understood your question correctly. Nir ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] Latest 3.5 - No Enable Gluster Service check box.
Latest 3.5 install as of today, I already have gluster up and running on all host and I am trying to import my gluster volume into ovirt. When I go to edit the cluster I have Enable Gluster Service but there is no check box next to it to check or uncheck. If I look at the host under Hosts, I do see: GlusterFS Version: glusterfs-3.5.2-1.el6 nathan stratton | vp technology | broadsoft, inc | +1-240-404-6580 | www.broadsoft.com ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] [RFI] oVirt 3.6 Planning
On 09/12/2014 04:45 PM, Federico Alberto Sayd wrote: On 12/09/14 09:55, Jakub Bittner wrote: ISO upload over web UI. +1, Is it so hard to implement such feature? well, the tricky part is web ui access the engine, which doesn't access the storage, rather the hosts do, so you need to stream it via the engine. good news are vdsm has now upload/download api's which should hopefully pave the way for this to materialize. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] [RFI] oVirt 3.6 Planning
On 09/12/2014 04:53 PM, Sven Kieske wrote: +1! I really would like to see a fast working way to create, manage and use ceph in ovirt. I don't know if implementation through openstack component is enough, because you will need more components then (keystone for auth e.g.). you would need to setup cinder with noauth or keystone, yes. similar to glance and neutron current support. neutron virtual appliance in 3.5 should have neutron with a local keystone already configured out of the box iirc On 12/09/14 15:34, Cédric Buot de l'Epine wrote: Dunno if CephFS efforts will be sustainable against Glusterfs... I'm not sure if a posix fs is the best provider for blocks. I would be pleased to have a direct form to create a Ceph pool for a datacenter, then provide librbd path for the guests (or only their size). ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] [RFI] oVirt 3.6 Planning
On 09/13/2014 03:32 PM, Mohyedeen Nazzal wrote: When VMs goes to unknown status, all control options are grayed, and the only way to fix this is to update the Database manually, it will be nice to be able to force the engine to poweroff the VM from admin portal. 'unknown' means the host is in non-responsive state. the correct approach is to correct the status of the host (stop via the power management options ('fence'), or 'confirm manual shutdown' of the host. either should release the 'unknown' status of the VM. for extreme cases (aka 'bugs'), there is a command line unlock utility to be used with caution. Thanks, Mohyedeen. On Sat, Sep 13, 2014 at 12:27 PM, Nir Soffer nsof...@redhat.com mailto:nsof...@redhat.com wrote: - Original Message - From: Sven Kieske s.kie...@mittwald.de mailto:s.kie...@mittwald.de To:users@ovirt.org mailto:users@ovirt.org Sent: Friday, September 12, 2014 4:53:46 PM Subject: Re: [ovirt-users] [RFI] oVirt 3.6 Planning +1! I really would like to see a fast working way to create, manage and use ceph in ovirt. I don't know if implementation through openstack component is enough, because you will need more components then (keystone for auth e.g.). Sure you will need more components, but we don't want to duplicate the effort, implementing ceph support directly in ovirt. Using Cinder we can reuse the existing ceph support, and gain also support for other Cinder drivers for free. On the host side, we will have of course ceph specific support, so you will be able to create ceph volumes directly, and use vdsm api to start vms using these volumes, if you want more direct support. I think this is the general long term approach; integrate openstack components instead of re-implementing the wheel. On 12/09/14 15:34, Cédric Buot de l'Epine wrote: Dunno if CephFS efforts will be sustainable against Glusterfs... I'm not sure if a posix fs is the best provider for blocks. I would be pleased to have a direct form to create a Ceph pool for a datacenter, then provide librbd path for the guests (or only their size). -- Mit freundlichen Grüßen / Regards Sven Kieske Systemadministrator Mittwald CM Service GmbH Co. KG Königsberger Straße 6 32339 Espelkamp T: +49-5772-293-100 tel:%2B49-5772-293-100 F: +49-5772-293-333 tel:%2B49-5772-293-333 https://www.mittwald.de Geschäftsführer: Robert Meyer St.Nr.: 331/5721/1033, USt-IdNr.: DE814773217, HRA 6640, AG Bad Oeynhausen Komplementärin: Robert Meyer Verwaltungs GmbH, HRB 13260, AG Bad Oeynhausen ___ Users mailing list Users@ovirt.org mailto:Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users ___ Users mailing list Users@ovirt.org mailto:Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users ___ 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: [ovirt-users] [RFI] oVirt 3.6 Planning
On 09/13/2014 03:38 PM, Mohyedeen Nazzal wrote: One more thing,, Being able to attach USB dongles from the admin portal... like the hostusb vdsm hook allows? Thanks, Mohyedeen. On Sat, Sep 13, 2014 at 3:32 PM, Mohyedeen Nazzal mohyedeen.naz...@gmail.com mailto:mohyedeen.naz...@gmail.com wrote: When VMs goes to unknown status, all control options are grayed, and the only way to fix this is to update the Database manually, it will be nice to be able to force the engine to poweroff the VM from admin portal. Thanks, Mohyedeen. On Sat, Sep 13, 2014 at 12:27 PM, Nir Soffer nsof...@redhat.com mailto:nsof...@redhat.com wrote: - Original Message - From: Sven Kieske s.kie...@mittwald.de mailto:s.kie...@mittwald.de To:users@ovirt.org mailto:users@ovirt.org Sent: Friday, September 12, 2014 4:53:46 PM Subject: Re: [ovirt-users] [RFI] oVirt 3.6 Planning +1! I really would like to see a fast working way to create, manage and use ceph in ovirt. I don't know if implementation through openstack component is enough, because you will need more components then (keystone for auth e.g.). Sure you will need more components, but we don't want to duplicate the effort, implementing ceph support directly in ovirt. Using Cinder we can reuse the existing ceph support, and gain also support for other Cinder drivers for free. On the host side, we will have of course ceph specific support, so you will be able to create ceph volumes directly, and use vdsm api to start vms using these volumes, if you want more direct support. I think this is the general long term approach; integrate openstack components instead of re-implementing the wheel. On 12/09/14 15:34, Cédric Buot de l'Epine wrote: Dunno if CephFS efforts will be sustainable against Glusterfs... I'm not sure if a posix fs is the best provider for blocks. I would be pleased to have a direct form to create a Ceph pool for a datacenter, then provide librbd path for the guests (or only their size). -- Mit freundlichen Grüßen / Regards Sven Kieske Systemadministrator Mittwald CM Service GmbH Co. KG Königsberger Straße 6 32339 Espelkamp T: +49-5772-293-100 tel:%2B49-5772-293-100 F: +49-5772-293-333 tel:%2B49-5772-293-333 https://www.mittwald.de Geschäftsführer: Robert Meyer St.Nr.: 331/5721/1033, USt-IdNr.: DE814773217, HRA 6640, AG Bad Oeynhausen Komplementärin: Robert Meyer Verwaltungs GmbH, HRB 13260, AG Bad Oeynhausen ___ Users mailing list Users@ovirt.org mailto:Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users ___ Users mailing list Users@ovirt.org mailto:Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users ___ 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: [ovirt-users] [RFI] oVirt 3.6 Planning
Le 13/09/2014 20:38, Nir Soffer a écrit : - Original Message - From: Nicolas Ecarnot nico...@ecarnot.net To: Nir Soffer nsof...@redhat.com Cc: users@ovirt.org, Federico Simoncelli fsimo...@redhat.com Sent: Saturday, September 13, 2014 9:24:55 PM Subject: Re: [ovirt-users] [RFI] oVirt 3.6 Planning Le 13/09/2014 17:34, Nir Soffer a écrit : For God's sake, please, prevent secondary storage domains (anything other than the master, especially iso and export) to completely block the whole thing when being unavailable. To many users suffered from this in the last two years I participate at this mailing list (including me). I think this is true for ISO and export domains for some time. In 3.4 at least, this is still true and is a problem. At least, if we can get rid of this and just mark these iso and export domains as unusable, and still do every other non-related operations, this would be very useful. It will not work for data domains - if you have vms with disks on what you call secondary data domain, how would you migrate these vms to a host that cannot see the secondary domain? In 3.6, there will be no master domain, so any data domain will be important as any other data domain. Maybe what do you like is to have more control on which domains are critical, and which are not. A domain which you mark as non-critical, or secondary, will not cause the host to become non-operational when the host cannot see this domain. So you would not be able to migrate some vms to a host that cannot see the secondary domain, but since *you* marked it as secondary, it is not a problem for you. What do you think? This proposition is perfect. I don't know whether being able to mark some storage domain as more precious than other will be useful, when their absence won't be a blocker anymore? The idea is to let you define which storage domains are required and which are not, lets call them optional. If a required storage domain cannot be seen, a host be become non-operational. If an optional storage domain cannot be seen, you will get a warning but the host will function normally. If you try to migrate a vm to a host which cannot see the storage domain used by that vm, the operation will fail. When a required storage domain is down, you would be able to change it to optional, and continue to work with the other domains in degraded mode. Some vms will not able to run, but other vms that do not depend on problem domain will not be affected. I hope that I understood your question correctly. Nir You did, and the features you described seem great. Can't wait to see them released! -- Nicolas Ecarnot ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] [RFI] oVirt 3.6 Planning
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 13.09.2014 22:25, Itamar Heim wrote: like the hostusb vdsm hook allows? is this installed by default, yet? furthermore I think you need to enable it via engine-config(custom parameters), don't you? I guess what is asked for is something working out of the box. kind regards Sven -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.22 (MingW32) iQGcBAEBAgAGBQJUFNv1AAoJEAq0kGAWDrqlNpgMAMgPo9QqH0aiqE0bMsmvXDbM R/gbK2+OteDGKgA7m4nvUIw5dzQ1X5a07+q51JoSRaUrJ3MxbA/y2lni9LG6ObyM 57pY3lwWqn91bS+dwo9bGaDKgdWq3HoHPfo+Wz7LfoHSACpeOgu+j6kWtgm3WcLm G3tST1fncLXnWPTQ6SkXaKo5N8G68NrjR2Ih8pFKu+GfeP+0MredKt3d+yHJRrjV 9ZtZvsvd4INDKOfVvJmR8MA0NvTSkFvLSBHvJ9gG3k0Umn/htdn9sPgS41Kghz2x NV1YX8pr4PmJjanGhFaVAYw2v367tVMgcewa21zZ162Wzh7nzTNOJAtMIfQn+hSG 2EGaw0NLFm0grpjITXCp24G1GwVUf3FR2dxt+GdsQ9upEs5605SMlknT6BLmtoJ8 8LCArwEfZ0UgC4T6qOWp7+Sbkcl124qg7abupewL8usKtmnQV00/8pIusx1mtypj 4BDAE5Cq4e+U8hiB9nlGJzo1IFnyBszMOQQLunKGow== =fv56 -END PGP SIGNATURE- ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] [RFI] oVirt 3.6 Planning
My vote is for storage load balancing/scheduler. Vmware's Vcenter has the concept of a 'storage cluster'. It's essentially a logical storage device. When you configure hosts/vms to use this device vcenter then works out which of the actual storage devices underneath this logical device is will send the storage requests to. This works out as a basic form of load balacing by alternating where the storage for new vms are created. This isn't particularly amazing, but what it does allow - with the highest end vcenter licensing anyway - is what Vmware calls 'storage distributed resource scheduling'. Much like we already have the ability to have a scheduler that moves the execution of vms around on hosts based on load, this does the same thing for the storage component of VM. Imagine having two configured storage locations under a 'storage cluster' and then having the ability to put one of the storage locations into 'maintenance mode'. The storage scheduler would then 'live storage migrate' all the storage for vms over to the other storage location. This would then allow the first storage location to be taken down for maintenance. This approach also allows storage to scale over time as more is added. The 'storage scheduler' can take inputs such as latency etc into account and manage the load across the 'storage cluster' to balance things out and make smart decisions so that the avaialble storage is utilized as best as it can be (ie: not overloading one storage location while the other location is mainly idle). I've done a bit a searching to see where Ovirt might be up to in this regard and what I've found seems to indicate that we are not anywhere near this capability just yet. An important prerequisite is having the hosts able to actually do a live storage migration. EL7 based hosts under ovirt 3.5 have this, as have Fedora 19 and 20 hosts. If the decision is made to use qemu-kvm-rhev on EL6 hosts - as has been talked about recently - then the host requirement for supporting live storage migration will be met. This then allows the idea of a storage scheduler to be futher considered. I think this is an important step in reaching feature parity with Vmware's vcenter product, and removes a key reason ovirt/rhev can't be considered in some instances. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] lost connectivty to the ovirt-engine after adding host
Hi, I am working on a proof of concept using using oVirt and ManageIQ. I have a CentOS6.5 where I deployed libvirt, I create a VM where I deployed oVirt 3.4. I was trying to add the host where the VM with oVirt resides. After much troubleshooting I was able to add the host but now the oVirt VM is owned by a user vdsm which I dont know the password for, so I can't start the VM with oVirt. This is the result of running virsh list $virsh list Please enter your authentication name: vdsm@ovirt Please enter your password: error: Failed to reconnect to the hypervisor error: no valid connection error: authentication failed: authentication failed I added another user with the command saslpasswd2 -a libvirt [username] I used that username and password to connect and to do virsh list -all That shows the VM but If i try to start it it gives me an error like follows error: Failed to start domain Cent6.5 error: internal error Process exited while reading console log output: char device redirected to /dev/pts/1 qemu-kvm: -drive file=/ovirt/test/virtlive01/miq_demo/cent01.img,if=none,id=drive-virtio-disk0,format=raw,cache=none: could not open disk image /ovirt/test/virtlive01/miq_demo/cent01.img: Permission denied Any ideas how can start that VM? is there a default password for the user vdsm ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users