Re: [Users] BSTRAP component='CreateConf' status='FAIL' message='Basic configuration failed to import default values'
On 07/04/2012 07:50 AM, Itamar Heim wrote: I have made some headway, managed to add the host into it's own cluster using ovirt-engine-cli executing: # create host --address xxx.xxx.xxx.xxx --cluster-id 99408929-82cf-4dc7-a532-9d998063fa95 --name Default --root_password Password --reboot_after_installation 0 But I haven't had much luck setting SPM -1. Would you mind showing me the exact syntax, please? michael? just add this option '--storage_manager-priority -1' to 'create' command, or use it for updating existent host. to see all available options for host creation do: help create host -- Michael Pasternak RedHat, ENG-Virtualization RD ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Ovirt and gluster storage (two servers in a cluster)
On 07/04/2012 12:36 AM, Robert Middleswarth wrote: On 07/04/2012 12:20 AM, зоррыч wrote: In what may be the reason for my situation? I had the problem using Fedora 17 using native posix fs mode. In your blog, I have not found any articles on ovirt. Last post by date it11/17/2011 I am still finishing up my blog post I haven't posted it yet It is still a draft. Getting closer. I am waking though the steps in a fresh install to make sure I am not missing any steps. In any case, if my situation has been reproduced and you - a sign of a bug that needs to be identified and transferred to the correction of the developers. When it was happening to me it looked like a locking problem. Since I was having other problems with Fedora 17 I rebuilt everything under CentOS 6.2 with zero problems. However I learned a few little tricks that might have helped make it not happen. In what configuration you have installed gluster? Both Posix FS (Native mode) and NFS using localhost:/volume format over a 1G network. Its performance depends on the choice of operating modes. In both cases provisioning was really slow. I found the NFS ran better of the two from research it was likely because NFS shares are by default cached native gluster fs isn't. However if you are using Fedora 17 with the latest kernel and it is stable for you Posix FS is likely much faster since fuse supports direct IO now. Also a faster network could make a real difference. Thanks for your help Thanks Robert *From:*Robert Middleswarth [mailto:rob...@middleswarth.net] *Sent:* Wednesday, July 04, 2012 6:15 AM *To:* зоррыч *Cc:* users@ovirt.org *Subject:* Re: [Users] Ovirt and gluster storage (two servers in a cluster) I am writing a blog post on the steps I did to get it all working. I have done it 3 times without issue. I am doing it a 4th time well writing I had a lot of problems with Fedora 17 but that was because of Fedora 17 not oVirt. Once I switched over to CentOS 6.2 and his build #8 everything just started working and working very well. On CentOS 6.2 I found that NFS is faster then native mode because of Caching over my gigabyte network but both are slow well provisioning. Thanks Robert On 07/03/2012 06:07 PM, зоррыч wrote: I use Scientific linux,and have already installed several times ovirt and vdsm, but the problem is reproduced. I sent the logs vdsm, but they do not specify any error. How do I know what is going wrong? I would be grateful for any ideas. I was just able to repeat the issue. If you only have one node active it will activate and work fine. But if you have 2 or more hosts / nodes it will just round robin though the hosts with each host contending on each round. I don't have that problem with NFS shares. Thanks Robert *From:*Robert Middleswarth [mailto:rob...@middleswarth.net] *Sent:* Wednesday, July 04, 2012 1:31 AM *To:* зоррыч *Cc:* users@ovirt.org mailto:users@ovirt.org *Subject:* Re: [Users] Ovirt and gluster storage (two servers in a cluster) I had the same problem with Fedora 17. I killed everything and started from scratch using centos and it works fine. I am wonder if there is the an issue with the Direct IO support that was added recently to Fedora 17. Thanks Robert On 07/03/2012 05:15 PM, зоррыч wrote: I can not mount a volume cluster with two servers. In a separate node activation, gluster volume installed successfully. For simultaneous operation of two servers with gluster volume ovirt constantly switches SPM, without giving an error. *From:*Robert Middleswarth [mailto:rob...@middleswarth.net] *Sent:* Wednesday, July 04, 2012 12:00 AM *To:* зоррыч *Cc:* users@ovirt.org mailto:users@ovirt.org *Subject:* Re: [Users] Ovirt and gluster storage (two servers in a cluster) Are you having problems creating a Volume or mounting the volume? Thanks Robert On 07/03/2012 03:56 PM, зоррыч wrote: I've updated ovirt and vdsm to the latest test version (git repository). But the situation continues to be repeated. What am I doing wrong? How do I find what is wrong? *From:*users-boun...@ovirt.org mailto:users-boun...@ovirt.org [mailto:users-boun...@ovirt.org] *On Behalf Of *?? *Sent:* Wednesday, June 27, 2012 6:19 PM *To:* rob...@middleswarth.net mailto:rob...@middleswarth.net *Cc:* users@ovirt.org mailto:users@ovirt.org *Subject:* Re: [Users] Ovirt and gluster storage (two servers in a cluster) logs in the attachment *From:*users-boun...@ovirt.org mailto:users-boun...@ovirt.org [mailto:users-boun...@ovirt.org] mailto:[mailto:users-boun...@ovirt.org] *On
Re: [Users] Ovirt and gluster storage (two servers in a cluster)
On 07/04/2012 12:18 PM, Robert Middleswarth wrote: I was just able to repeat the issue. If you only have one node active it will activate and work fine. But if you have 2 or more hosts / nodes it will just round robin though the hosts with each host contending on each round. I don't have that problem with NFS shares. Does mounting the volume fail when you have two nodes? -Vijay ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[Users] What does this small icon mean in engine?
Hi, Please check the attachment to see the information about my two disks. And both of disk2 and disk3 do have a small icon like a hand in the fourth column. The first column is Alias and the second is ID. There is no description in the third and the fourth column. What dose the small icon mean? -- Shu Ming shum...@linux.vnet.ibm.com IBM China Systems and Technology Laboratory attachment: junk.JPG___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Setting up test nodes
On 03/07/12 21:27, Itamar Heim wrote: On 07/03/2012 03:48 PM, Martin Kletzander wrote: Hi everyone, TL;DR: how do I setup a software node (faqemu) with apps from source? - do you wnat to setup an ovirt-node, or just install vdsm on plain fedora 17? sounds like the latter, but if you want to use fake-qemu, ovirt-node is much harder than just using a fedora 17. I'm new to the oVirt world and coming from the lower layer, I must admin I feel kind of confused. I'm trying to create a lab with ovirt-engine (up and running), some ovirt-node (with sw qemu) and ovirt-node with all the stack running on git sources (qemu, libvirt, vdsm and ovirt-node). The problem is that I only tried this before with ovirt-node ISO image and I don't know how can I step in there and work with the system underneath. The whole point of this is to help speed up problem-solving in the future for problems related to mostly libvirt. Having this available should help us a lot. My main question is: What is the proper way to setup a node from standard fedora installation (not ISO) and having software-emulated qemu machine there? for using plain fedora 17 as a guest, you just need to: option 1: - yum install vdsm vdsdsm-hook-faqemu - vi /etc/vdsm/vdsm.conf and set vars.fake_kvm_support=True - now simply add the host from web admin (hosts--add host) note: you may need this patch if not in your version of vdsm already: http://gerrit.ovirt.org/#/c/5611/ option 2: if your host is fedora 16 and above, just use nested virtualization, and your virtual host would behave like a normal one (i still need to try this one out) i.e., just add the virtual host from web admin (hosts--add host) I assume another step would be needed here (at least configuring the guest in libvirt to have nested virtualization), but i haven't tried this one yet to know what it is. Once successful, it would be great if you could document your process in oVirt's wiki, sharing your experience with other users ;) Thanks in advance for any tips and have a nice day, Martin -- /d Willyoupleasehelpmefixmykeyboard?Thespacebarisbroken! ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Unable to run VM with Unable to create cgroup error message
Sorry, I almost always forget the links :)
[1] http://libvirt.org/logging.html
Martin
P.S.: don't hesitate to send other info and ask, I'd be happy to help,
although I'm going away for week and a half.
On 07/04/2012 10:12 AM, Martin Kletzander wrote:
Hi,
the only problem I could imagine is that there is really some directory
missing, but you said it is there. Fromt his point I see two things you
can try:
1) run libvirt with debugging output enabled [1]
2) try to create the directories manually
Hopefully it'll give us more information on what's wrong.
Martin
On 07/03/2012 11:55 AM, xuejie chen wrote:
Hi everyone,
I created a Desktop VM1 with all default settings and added a 2G Disk
to the VM1 in WebAdmin.
But when I run the vm, it returns failure with the follow error
message in WebAdmin.
VM VM1 is down. Exit message: Unable to create cgroup for VM1
My DateCenter contains one host(with OS is CentOS 6, VDSM version is
4.9.6 and libvirt version is 0.9.4) and one NFS domain.
And The cgroup Directory have already exits on the host.
But why cannot I run the VM1?
There are some error message in VDSM log
File /usr/lib64/python2.6/site-packages/libvirt.py, line 2087, in createXML
if ret is None: raise libvirtError('virDomainCreateXML() failed',
conn=self)
libvirtError: Unable to create cgroup for VM1: No such file or directory
--
There are some error message in libvirt log
-
error : virNetClientProgramDispatchError:170 : Unable to create cgroup
for VM1: No such file or directory
-
There are log files in attachment
Best wishes,
Xuejie Chen
___
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: [Users] Setting up test nodes
On 07/04/2012 10:46 AM, Doron Fediuck wrote: On 03/07/12 21:27, Itamar Heim wrote: On 07/03/2012 03:48 PM, Martin Kletzander wrote: Hi everyone, TL;DR: how do I setup a software node (faqemu) with apps from source? - do you wnat to setup an ovirt-node, or just install vdsm on plain fedora 17? sounds like the latter, but if you want to use fake-qemu, ovirt-node is much harder than just using a fedora 17. Frankly, I don't see the difference in there that much. The basic point in this is -- have a visible node in ovirt-engine on which I can experiment with (mainly) libvirt. At first I though I need to have the 'ovirt-node' package installed but it doesn't seem it's needed now. I searched through the documentation but I haven't found a lot related to this particular use case. I'm new to the oVirt world and coming from the lower layer, I must admin I feel kind of confused. I'm trying to create a lab with ovirt-engine (up and running), some ovirt-node (with sw qemu) and ovirt-node with all the stack running on git sources (qemu, libvirt, vdsm and ovirt-node). The problem is that I only tried this before with ovirt-node ISO image and I don't know how can I step in there and work with the system underneath. The whole point of this is to help speed up problem-solving in the future for problems related to mostly libvirt. Having this available should help us a lot. My main question is: What is the proper way to setup a node from standard fedora installation (not ISO) and having software-emulated qemu machine there? for using plain fedora 17 as a guest, you just need to: option 1: - yum install vdsm vdsdsm-hook-faqemu - vi /etc/vdsm/vdsm.conf and set vars.fake_kvm_support=True - now simply add the host from web admin (hosts--add host) note: you may need this patch if not in your version of vdsm already: http://gerrit.ovirt.org/#/c/5611/ I'll try this after resolving few other errors unrelated to oVirt. I though there is some more complicated way through swamps and dragon lair =) Thanks for showing me the right way! option 2: if your host is fedora 16 and above, just use nested virtualization, and your virtual host would behave like a normal one (i still need to try this one out) i.e., just add the virtual host from web admin (hosts--add host) I wouldn't want to try that, kvm is still not that stable, people say the guests get stuck after some time and I don't really need anything running in these machines. In case you'll try that, good luck. I assume another step would be needed here (at least configuring the guest in libvirt to have nested virtualization), but i haven't tried this one yet to know what it is. Once successful, it would be great if you could document your process in oVirt's wiki, sharing your experience with other users ;) I'll try to summarize that in case I'll succeed, but now everything seems very new to me (I mean I don't even know where to click in the engine's administration portal). Thanks in advance for any tips and have a nice day, Martin ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Ovirt and gluster storage (two servers in a cluster)
On 07/04/2012 12:58 PM, Robert Middleswarth wrote: On 07/04/2012 03:15 AM, Vijay Bellur wrote: On 07/04/2012 12:18 PM, Robert Middleswarth wrote: I was just able to repeat the issue. If you only have one node active it will activate and work fine. But if you have 2 or more hosts / nodes it will just round robin though the hosts with each host contending on each round. I don't have that problem with NFS shares. Does mounting the volume fail when you have two nodes? -Vijay No. It will mount and activate for a few seconds then the 2nd host will start contending for SPM. Will become the SPM host then the 1st host will start contending rise and repeat the data center is never up for more then a few seconds. If I put all hosts but one into maintenance then the data center will become active and will work fine including allowing me to start hosts but if I take the host out of maintenance then all the hosts will fight for SPM and the datacenter will never become active. Can you please send across vdsm, engine and glusterfs logs? Thanks, Vijay ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Ovirt and gluster storage (two servers in a cluster)
On 07/04/2012 10:28 AM, Robert Middleswarth wrote: On 07/04/2012 03:15 AM, Vijay Bellur wrote: On 07/04/2012 12:18 PM, Robert Middleswarth wrote: I was just able to repeat the issue. If you only have one node active it will activate and work fine. But if you have 2 or more hosts / nodes it will just round robin though the hosts with each host contending on each round. I don't have that problem with NFS shares. Does mounting the volume fail when you have two nodes? -Vijay No. It will mount and activate for a few seconds then the 2nd host will start contending for SPM. Will become the SPM host then the 1st host will start contending rise and repeat the data center is never up for more then a few seconds. If I put all hosts but one into maintenance then the data center will become active and will work fine including allowing me to start hosts but if I take the host out of maintenance then all the hosts will fight for SPM and the datacenter will never become active. Thanks Robert https://bugzilla.redhat.com/show_bug.cgi?id=835949 ? Y. ___ 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] Removal of network failing from DataCenter.
Hi Folks,
I was trying to remove networks from the Cluster and
Data-Center using the python bindings for ovirt. I could remove the network
from the cluster but it fails to remove the network from the datacenter
quitting with the below mentioned traceback :
* Traceback (most recent call last):*
* File netdel.py, line 25, in module*
*Netwk.delete()*
* File
/usr/lib/python2.6/site-packages/ovirt_engine_sdk-1.5_SNAPSHOT-py2.6.egg/ovirtsdk/infrastructure/brokers.py,
line 1889, in delete*
*headers={'Content-type':None})*
* File
/usr/lib/python2.6/site-packages/ovirt_engine_sdk-1.5_SNAPSHOT-py2.6.egg/ovirtsdk/infrastructure/proxy.py,
line 36, in delete*
*return self.request('DELETE', url, body, headers)*
* File
/usr/lib/python2.6/site-packages/ovirt_engine_sdk-1.5_SNAPSHOT-py2.6.egg/ovirtsdk/infrastructure/proxy.py,
line 52, in request*
*conn=self.getConnectionsPool().getConnection())*
* File
/usr/lib/python2.6/site-packages/ovirt_engine_sdk-1.5_SNAPSHOT-py2.6.egg/ovirtsdk/infrastructure/proxy.py,
line 62, in __doRequest*
*raise RequestError, response*
*ovirtsdk.infrastructure.errors.RequestError:*
*status: 400*
*reason: Bad Request*
*detail: [Cannot remove Network. Network is being used by at least one
Cluster.]*
However, I am able to remove it when I try from the Manager UI. Below is
the code I used to test.I am not able to figure-out if it is a know bug or
a logical fault on my side somewhere ? [All the VMs are shutdown and no VM
is attached to network that I am deleting.]
*from ovirtsdk.api import API*
*from ovirtsdk.xml import params*
*
*
*
*
*ovirt_uri = http://ovirtserver.xyz.com:8080/api*
*ovirt_username = admin@internal*
*ovirt_password = **
*
*
*ovirtAPI = API(url=ovirt_uri, username=ovirt_username,
password=ovirt_password)*
*
*
*
*
*## removing the network from the cluster.*
*Netw=ovirtAPI.clusters.get(Default).networks.get(testnet58)*
*print Netw.name + : + Netw.id*
*Netw.delete()*
*
*
*## removing the network from the Datacenter.*
*Netwk=ovirtAPI.networks.get(testnet58)*
*print Netwk.name + : + Netwk.id*
*Netwk.delete()*
--
Regards,
Rahul
===
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Unable to run VM with Unable to create cgroup error message
- Original Message -
From: Martin Kletzander mklet...@redhat.com
To: xuejie chen xuejieche...@gmail.com
Cc: users@ovirt.org
Sent: Wednesday, July 4, 2012 4:12:07 AM
Subject: Re: [Users] Unable to run VM with Unable to create cgroup error
message
Hi,
the only problem I could imagine is that there is really some
directory
missing, but you said it is there. Fromt his point I see two things
you
can try:
1) run libvirt with debugging output enabled [1]
2) try to create the directories manually
Is the cgroupfs mounted?
Hopefully it'll give us more information on what's wrong.
Martin
On 07/03/2012 11:55 AM, xuejie chen wrote:
Hi everyone,
I created a Desktop VM1 with all default settings and added a 2G
Disk
to the VM1 in WebAdmin.
But when I run the vm, it returns failure with the follow error
message in WebAdmin.
VM VM1 is down. Exit message: Unable to create cgroup for VM1
My DateCenter contains one host(with OS is CentOS 6, VDSM version
is
4.9.6 and libvirt version is 0.9.4) and one NFS domain.
And The cgroup Directory have already exits on the host.
But why cannot I run the VM1?
There are some error message in VDSM log
File /usr/lib64/python2.6/site-packages/libvirt.py, line 2087, in
createXML
if ret is None: raise libvirtError('virDomainCreateXML()
failed', conn=self)
libvirtError: Unable to create cgroup for VM1: No such file or
directory
--
There are some error message in libvirt log
-
error : virNetClientProgramDispatchError:170 : Unable to create
cgroup
for VM1: No such file or directory
-
There are log files in attachment
Best wishes,
Xuejie Chen
___
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: [Users] BSTRAP component='CreateConf' status='FAIL' message='Basic configuration failed to import default values'
On 07/04/2012 07:50 AM, Itamar Heim wrote: I have made some headway, managed to add the host into it's own cluster using ovirt-engine-cli executing: # create host --address xxx.xxx.xxx.xxx --cluster-id 99408929-82cf-4dc7-a532-9d998063fa95 --name Default --root_password Password --reboot_after_installation 0 But I haven't had much luck setting SPM -1. Would you mind showing me the exact syntax, please? michael? just add this option '--storage_manager-priority -1' to 'create' command, or use it for updating existent host. Sure thing. Could you also give an example command on how you modify that property on an already added host? /Karli to see all available options for host creation do: help create host -- Michael Pasternak RedHat, ENG-Virtualization RD ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] BSTRAP component='CreateConf' status='FAIL' message='Basic configuration failed to import default values'
On 07/04/2012 04:23 PM, Karli Sjöberg wrote: On 07/04/2012 07:50 AM, Itamar Heim wrote: I have made some headway, managed to add the host into it's own cluster using ovirt-engine-cli executing: # create host --address xxx.xxx.xxx.xxx --cluster-id 99408929-82cf-4dc7-a532-9d998063fa95 --name Default --root_password Password --reboot_after_installation 0 But I haven't had much luck setting SPM -1. Would you mind showing me the exact syntax, please? michael? just add this option '--storage_manager-priority -1' to 'create' command, or use it for updating existent host. Sure thing. Could you also give an example command on how you modify that property on an already added host? update host myhost --storage_manager-priority -1 /Karli to see all available options for host creation do: help create host -- Michael Pasternak RedHat, ENG-Virtualization RD -- Michael Pasternak RedHat, ENG-Virtualization RD ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Unable to run VM with Unable to create cgroup error message
On 07/04/2012 02:36 PM, Andrew Cathrow wrote: - Original Message - From: Martin Kletzander mklet...@redhat.com To: xuejie chen xuejieche...@gmail.com Cc: users@ovirt.org Sent: Wednesday, July 4, 2012 4:12:07 AM Subject: Re: [Users] Unable to run VM with Unable to create cgroup error message Hi, the only problem I could imagine is that there is really some directory missing, but you said it is there. Fromt his point I see two things you can try: 1) run libvirt with debugging output enabled [1] 2) try to create the directories manually Is the cgroupfs mounted? I checked and libvirt should complain if it is not mounted. But to be sure, is it? ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[Users] oVirt Weekly Meeting - Delayed to 2012-07-05
Since today is 04-July, all US people are un-available. The meeting will take place tomorrow, same place, same hour. Thanks, Ofer Schreiber oVirt Release Manager ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] BSTRAP component='CreateConf' status='FAIL' message='Basic configuration failed to import default values'
On 07/04/2012 04:23 PM, Karli Sjöberg wrote: On 07/04/2012 07:50 AM, Itamar Heim wrote: I have made some headway, managed to add the host into it's own cluster using ovirt-engine-cli executing: # create host --address xxx.xxx.xxx.xxx --cluster-id 99408929-82cf-4dc7-a532-9d998063fa95 --name Default --root_password Password --reboot_after_installation 0 But I haven't had much luck setting SPM -1. Would you mind showing me the exact syntax, please? michael? just add this option '--storage_manager-priority -1' to 'create' command, or use it for updating existent host. Sure thing. Could you also give an example command on how you modify that property on an already added host? update host myhost --storage_manager-priority -1 Awesome, thanks man! /Karli /Karli to see all available options for host creation do: help create host -- Michael Pasternak RedHat, ENG-Virtualization RD -- Michael Pasternak RedHat, ENG-Virtualization RD ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Online snapshot
After looking into a qemu a bit, I believe the snapshot operation on the disk in oVirt is not a live snapshot, because the guest VM will be paused by qemu when the snapshot is undergoing.That is what qemu monitor command snapshot_blkdev does and VDSM depends on that command in low level.Most likely, the pausing time is non-trival for a normal snapshot. On 2012-6-29 0:37, Itamar Heim wrote: On 06/28/2012 09:20 AM, Eric Blake wrote: On 06/28/2012 07:07 AM, Dave Allan wrote: On Thu, Jun 28, 2012 at 06:34:25AM -0400, Itamar Heim wrote: On 06/28/2012 04:40 AM, Winfried de Heiden wrote: Hi all, It seems live snapshot is available in the Fedora 17 qemu-kvm. On my laptop with kvm, virt-manager, libvirt, virsh etc. I created a virtual machine and tried to make a live snapshot: virsh # snapshot-create RHEL55_NTP Domain snapshot 1340871802 created That's a system-checkpoint snapshot, because you didn't use --disk-only. When you say 'live snapshot' were you intending the fast operation of just a disk snapshot (via the 'transaction' or 'blockdev-snapshot-sync' monitor command) or the slow operation of a checkpoint (via the 'savevm' monitor command)? [root@kvm ~]# rpm -qf `which qemu-kvm` qemu-system-x86-1.0-17.fc17.x86_64 [wdh@laptop-winfried ~]$ rpm -qf `which qemu-kvm` qemu-system-x86-1.0-17.fc17.x86_64 This qemu does not support the 'transaction' monitor command, so libvirt has to fall back to the non-atomic 'blockdev-snapshot-sync', or else give up because of missing functionality. dave - who can help us elaborate on this libvirt error: reuse is not supported with this QEMU binary this is the relevant error coming from libvirt. maybe fedora 17 qemu-kvm does not have live snapshot support? Thread-306::ERROR::2012-06-27 12:46:40,178::libvirtvm::1782::vm.Vm::(snapshot) vmId=`ef978066-09c8-40aa-8e0b-4accfacc15bd`::Unable to take snapshot Traceback (most recent call last): File /usr/share/vdsm/libvirtvm.py, line 1780, in snapshot libvirt.VIR_DOMAIN_SNAPSHOT_CREATE_REUSE_EXT)) OK, so this is a case of VDSM driving a different use of the API than what you did with your virsh command line. VDSM is attempting to do the equivalent of 'virsh snapshot-create --disk-only --reuse-ext', but because your qemu doesn't support the 'transaction' command, libvirt can't make qemu do what VDSM wants. Basically, the 'blockdev-snapshot-sync' command in qemu 1.0 lacks the ability to reuse qcow2 metadata already present in the file pre-created by VDSM; this wasn't added until the 'transaction' command of qemu 1.1. ok, i guess the question is then when qemu 1.1 is expected in fedora ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users -- Shu Ming shum...@linux.vnet.ibm.com IBM China Systems and Technology Laboratory ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] What does this small icon mean in engine?
On 07/04/2012 10:25 AM, Shu Ming wrote: Hi, Please check the attachment to see the information about my two disks. And both of disk2 and disk3 do have a small icon like a hand in the fourth column. The first column is Alias and the second is ID. There is no description in the third and the fourth column. What dose the small icon mean? iirc, the hand means this disk has the 'shared' checkbox ___ 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] oVirt 3.1 and Glusterfs how-to
I wrote a basic how to for using glusterfs under oVirt. If you have any questions or you find mistakes please let me know. Thanks Robert http://www.middleswarth.net/content/installing-ovirt-31-and-glusterfs-using-either-nfs-or-posix-native-file-system ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Ovirt and gluster storage (two servers in a cluster)
On 07/04/2012 07:42 AM, Yaniv Kaul wrote: On 07/04/2012 10:28 AM, Robert Middleswarth wrote: On 07/04/2012 03:15 AM, Vijay Bellur wrote: On 07/04/2012 12:18 PM, Robert Middleswarth wrote: I was just able to repeat the issue. If you only have one node active it will activate and work fine. But if you have 2 or more hosts / nodes it will just round robin though the hosts with each host contending on each round. I don't have that problem with NFS shares. Does mounting the volume fail when you have two nodes? -Vijay No. It will mount and activate for a few seconds then the 2nd host will start contending for SPM. Will become the SPM host then the 1st host will start contending rise and repeat the data center is never up for more then a few seconds. If I put all hosts but one into maintenance then the data center will become active and will work fine including allowing me to start hosts but if I take the host out of maintenance then all the hosts will fight for SPM and the datacenter will never become active. Thanks Robert https://bugzilla.redhat.com/show_bug.cgi?id=835949 ? Y. That is clearly the issue I am seeing. Thanks Robert ___ 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] Unable to run VM with Unable to create cgroup error message
Hi, 2012/7/4 Martin Kletzander mklet...@redhat.com: On 07/04/2012 02:36 PM, Andrew Cathrow wrote: - Original Message - From: Martin Kletzander mklet...@redhat.com To: xuejie chen xuejieche...@gmail.com Cc: users@ovirt.org Sent: Wednesday, July 4, 2012 4:12:07 AM Subject: Re: [Users] Unable to run VM with Unable to create cgroup error message Hi, the only problem I could imagine is that there is really some directory missing, but you said it is there. Fromt his point I see two things you can try: 1) run libvirt with debugging output enabled [1] 2) try to create the directories manually Is the cgroupfs mounted? I checked and libvirt should complain if it is not mounted. But to be sure, is it? I do not know how to check it. where the cgroupfs should be mounted to? there are some cgroup directory infomation. -- #ll / . drwxr-xr-x.3 root root 0 Jul 2 16:56 cgroup # ls /cgroup cgroup.procs cpu.cfs_period_us cpu.cfs_quata_us there are user and group is root root. Is it because the libvirt donot have permission to operate the cgroup? Thanks a lot. xuejie chen. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Unable to run VM with Unable to create cgroup error message
2012/7/5 xuejie chen xuejieche...@gmail.com: Hi, 2012/7/4 Martin Kletzander mklet...@redhat.com: On 07/04/2012 02:36 PM, Andrew Cathrow wrote: - Original Message - From: Martin Kletzander mklet...@redhat.com To: xuejie chen xuejieche...@gmail.com Cc: users@ovirt.org Sent: Wednesday, July 4, 2012 4:12:07 AM Subject: Re: [Users] Unable to run VM with Unable to create cgroup error message Hi, the only problem I could imagine is that there is really some directory missing, but you said it is there. Fromt his point I see two things you can try: 1) run libvirt with debugging output enabled [1] 2) try to create the directories manually Is the cgroupfs mounted? I checked and libvirt should complain if it is not mounted. But to be sure, is it? I do not know how to check it. where the cgroupfs should be mounted to? I have confirmed the cgroupfs did not mounted. Thanks a lot. xuejie chen. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
