Re: [Users] Storage server
On 02/04/2014 03:10 PM, Elad Ben Aharon wrote: Actually, the best way to do it would be to create another storage domain and migrate all the VMs's disks to it (you can do it while the VMs are running), then you won't suffer from a down time. notice this will create a snapshot for your vms and may change them from raw to qcow. also, this not supported for shared disks. If you don't have the option to do so, the best way to do it would be to shut down the VMs, put the storage domain to 'Maintenance' and then perform the storage server network change. When the connection to the storage is fixed, activate the domain and resume the VMs manually. that's assuming the domain was added using dns and not ip address, and that the dns expiration/caching on all hosts will notice the new ip address. - Original Message - From: Koen Vanoppen vanoppen.k...@gmail.com To: Elad Ben Aharon ebena...@redhat.com, users@ovirt.org Sent: Tuesday, February 4, 2014 2:57:57 PM Subject: Re: [Users] Storage server For the moment, nothing is going on. The change is planned on Thursday. So, correct me if I'm wrong, you say that the fastest and easiest way is to suspend all the machines, make the change of the storage server, rediscover the iscsi server and then bring the vm's back up? Kind regards, Koen On Feb 4, 2014 1:47 PM, Elad Ben Aharon ebena...@redhat.com wrote: Hi, I suppose that those VMs are now in 'Paused' state, correct me if I'm wrong. After connectivity with storage comes back, all VMs which couldn't detect their disks should be resumed from 'Paused' to 'Up' state automatically. It shouldn't take them a long time to change their state (a few minutes). - Original Message - From: Koen Vanoppen vanoppen.k...@gmail.com To: users@ovirt.org Sent: Tuesday, February 4, 2014 2:32:46 PM Subject: [Users] Storage server Dear All, At work, the network guys have come to a discovery that they have made a mistake in their network configuration... The Storage server that we use as ISCSI target in oVirt, is in the wrong VLAN, so the wrong IP address... What is the best way to make sure that all of our vms (40) that are using this scsi storage will come up again after the storage server has changed his ip? What are the best steps to be taken...? We are using oVirt Engine Version: 3.3.1-2.el6 Kind Regards, Koen ___ 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] NAT configuration
It's actually step 5 :) Am 05.02.2014 07:50, schrieb Itamar Heim: (step 6 explains where the user/password for libvirt/virsh are. -- 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] ovirt engine and vdsm on same node
Can we install ovirt engine and vdsm on same node ? If so any down side ? -- *Hans Emmanuel* *NOthing to FEAR but something to FEEL..* ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] ovirt engine and vdsm on same node
Il 05/02/2014 09:18, Hans Emmanuel ha scritto: Can we install ovirt engine and vdsm on same node ? Sure, you can use all-in-one plugin for having both running on the same node. In 3.4.0 you'll be able also to run ovirt engine inside a VM using hosted-engine feature. If so any down side ? You'll have less resources available for VMs on the host running the engine. -- *Hans Emmanuel* /NOthing to FEAR but something to FEEL../ ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users -- Sandro Bonazzola Better technology. Faster innovation. Powered by community collaboration. See how it works at redhat.com ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] NAT configuration
Well, I just tested this. and I can't connect to virsh with this information. I guess the mentioned user vdsm@rhevh might not be the actual one used in 3.3.2 anymore? (mail is from 2012 and mentions rhev, so..) or can't libvirt manage multiple authenticated users? Because I registered my own using: saslpasswd2 -a libvirt USERNAME which still works. Am 05.02.2014 09:15, schrieb Sven Kieske: It's actually step 5 :) Am 05.02.2014 07:50, schrieb Itamar Heim: (step 6 explains where the user/password for libvirt/virsh are. -- 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
Re: [Users] ovirt engine and vdsm on same node
- Original Message - From: Sandro Bonazzola sbona...@redhat.com To: Hans Emmanuel hansemman...@gmail.com, users@ovirt.org Sent: Wednesday, February 5, 2014 10:22:28 AM Subject: Re: [Users] ovirt engine and vdsm on same node Il 05/02/2014 09:18, Hans Emmanuel ha scritto: Can we install ovirt engine and vdsm on same node ? Sure, you can use all-in-one plugin for having both running on the same node. In 3.4.0 you'll be able also to run ovirt engine inside a VM using hosted-engine feature. Also note that doing this manually, not using the all-in-one plugin, is unsupported and might cause problems (iirc _will_ cause). Meaning: Do _not_ do the following: install and setup engine, then add the host it's on as a host. If so any down side ? You'll have less resources available for VMs on the host running the engine. It's also a bit more complex, and generally considered suitable only for very small installations (just one machine) etc. With hosted-engine, even if you start with a single host, if you add one in the future, you'll be able to migrate the engine VM to the other host and have HA for it. -- Didi ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] ovirt engine and vdsm on same node
Thanks for your inputs . On Wed, Feb 5, 2014 at 2:04 PM, Yedidyah Bar David d...@redhat.com wrote: - Original Message - From: Sandro Bonazzola sbona...@redhat.com To: Hans Emmanuel hansemman...@gmail.com, users@ovirt.org Sent: Wednesday, February 5, 2014 10:22:28 AM Subject: Re: [Users] ovirt engine and vdsm on same node Il 05/02/2014 09:18, Hans Emmanuel ha scritto: Can we install ovirt engine and vdsm on same node ? Sure, you can use all-in-one plugin for having both running on the same node. In 3.4.0 you'll be able also to run ovirt engine inside a VM using hosted-engine feature. Also note that doing this manually, not using the all-in-one plugin, is unsupported and might cause problems (iirc _will_ cause). Meaning: Do _not_ do the following: install and setup engine, then add the host it's on as a host. If so any down side ? You'll have less resources available for VMs on the host running the engine. It's also a bit more complex, and generally considered suitable only for very small installations (just one machine) etc. With hosted-engine, even if you start with a single host, if you add one in the future, you'll be able to migrate the engine VM to the other host and have HA for it. -- Didi -- *Hans Emmanuel* *NOthing to FEAR but something to FEEL..* ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Quotas
Hi Maurice, With quota you can set a hard limit for the number of CPU consumed by the user, and by that limit the num of vms, other quota parameters (memory and storage) can be unlimited. Although these videos are for version 3.1 they are pretty useful, since there weren't any drastic change since then: http://www.youtube.com/watch?v=zazJ_fW05Qk http://www.youtube.com/watch?v=y8qUoRImimY you are more than welcome to send your progress/use-case/any other question :-) Thanks, Gilad. - Original Message - From: Maurice James midnightst...@msn.com To: users@ovirt.org Sent: Wednesday, February 5, 2014 3:47:51 AM Subject: [Users] Quotas How do I limit the number of VMs that a user can create using quotas? ___ 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] NAT configuration
On 5-2-2014 9:27, Sven Kieske wrote: Well, I just tested this. and I can't connect to virsh with this information. I guess the mentioned user vdsm@rhevh might not be the actual one used in 3.3.2 anymore? (mail is from 2012 and mentions rhev, so..) vdsm@ovirt seems to work :-) Joop ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Quotas
- Original Message - From: Maurice James midnightst...@msn.com To: users@ovirt.org Sent: Wednesday, February 5, 2014 3:47:51 AM Subject: [Users] Quotas How do I limit the number of VMs that a user can create using quotas? Hi Maurice. Quota supports the resource level (storage, CPU, RAM) rather than VM count. So please open an RFE with your needs so we'll be able to add it to the requested tasks. We can try to hack it using a scheduling filter, but I'd rather have a proper support for such a request. Doron ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] failed nestd vm only with spice and not vnc
Hello,
I revamp this thread putting a subject more in line with the real problem.
Previous thread subject was
unable to start vm in 3.3 and f19 with gluster
and began here on ovirt users mailing list:
http://lists.ovirt.org/pipermail/users/2013-September/016628.html
Now I updated all to final 3.3.3 and I see that the problem is here yet.
So now I have updated Fedora 19 hosts that are VMs (virtual hw version
9) inside vSphere infra version 5.1.
CPU of ESX host is E7-4870 and cluster in oVirt is defined as Intel
Nehalem Family
On oVirt host VM
[root@ovnode01 qemu]# rpm -q libvirt qemu-kvm
libvirt-1.0.5.9-1.fc19.x86_64
qemu-kvm-1.4.2-15.fc19.x86_64
[root@ovnode01 qemu]# uname -r
3.12.9-201.fc19.x86_64
flags of cpuinfo:
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
mca cmov pat pse36 clflush dts mmx fxsr sse sse2 ss syscall nx rdtscp
lm constant_tsc arch_perfmon pebs bts nopl xtopology tsc_reliable
nonstop_tsc aperfmperf pni monitor vmx ssse3 cx16 sse4_1 sse4_2 x2apic
popcnt lahf_lm ida arat epb dtherm tpr_shadow vnmi ept vpid
[root@ovnode01 ~]# vdsClient -s localhost getVdsCapabilities
HBAInventory = {'FC': [], 'iSCSI': [{'InitiatorName':
'iqn.1994-05.com.redhat:6344c23973df'}]}
ISCSIInitiatorName = 'iqn.1994-05.com.redhat:6344c23973df'
bondings = {'bond0': {'addr': '',
'cfg': {},
'hwaddr': '32:5c:6a:20:cd:21',
'ipv6addrs': [],
'mtu': '1500',
'netmask': '',
'slaves': []}}
bridges = {'ovirtmgmt': {'addr': '192.168.33.41',
'cfg': {'BOOTPROTO': 'none',
'DEFROUTE': 'yes',
'DELAY': '0',
'DEVICE': 'ovirtmgmt',
'GATEWAY': '192.168.33.15',
'IPADDR': '192.168.33.41',
'NETMASK': '255.255.255.0',
'NM_CONTROLLED': 'no',
'ONBOOT': 'yes',
'STP': 'no',
'TYPE': 'Bridge'},
'gateway': '192.168.33.15',
'ipv6addrs': ['fe80::250:56ff:fe9f:686b/64'],
'ipv6gateway': '::',
'mtu': '1500',
'netmask': '255.255.255.0',
'ports': ['eth0', 'vnet1'],
'stp': 'off'},
'vlan172': {'addr': '',
'cfg': {'DEFROUTE': 'no',
'DELAY': '0',
'DEVICE': 'vlan172',
'NM_CONTROLLED': 'no',
'ONBOOT': 'yes',
'STP': 'no',
'TYPE': 'Bridge'},
'gateway': '0.0.0.0',
'ipv6addrs': ['fe80::250:56ff:fe9f:3b86/64'],
'ipv6gateway': '::',
'mtu': '1500',
'netmask': '',
'ports': ['ens256.172', 'vnet0'],
'stp': 'off'}}
clusterLevels = ['3.0', '3.1', '3.2', '3.3']
cpuCores = '4'
cpuFlags =
'fpu,vme,de,pse,tsc,msr,pae,mce,cx8,apic,sep,mtrr,pge,mca,cmov,pat,pse36,clflush,dts,mmx,fxsr,sse,sse2,ss,syscall,nx,rdtscp,lm,constant_tsc,arch_perfmon,pebs,bts,nopl,xtopology,tsc_reliable,nonstop_tsc,aperfmperf,pni,monitor,vmx,ssse3,cx16,sse4_1,sse4_2,x2apic,popcnt,lahf_lm,ida,arat,epb,dtherm,tpr_shadow,vnmi,ept,vpid,model_Nehalem,model_Conroe,model_coreduo,model_core2duo,model_Penryn,model_n270'
cpuModel = 'Intel(R) Xeon(R) CPU E7- 4870 @ 2.40GHz'
cpuSockets = '4'
cpuSpeed = '2394.000'
cpuThreads = '4'
emulatedMachines = ['pc',
'q35',
'isapc',
'pc-0.10',
'pc-0.11',
'pc-0.12',
'pc-0.13',
'pc-0.14',
'pc-0.15',
'pc-1.0',
'pc-1.1',
'pc-1.2',
'pc-1.3',
'none']
guestOverhead = '65'
hooks = {}
kvmEnabled = 'true'
lastClient = '127.0.0.1'
lastClientIface = 'lo'
management_ip = '0.0.0.0'
memSize = '16050'
Re: [Users] NAT configuration
I can confirm that vdsm@ovirt does work. However, I have the strong feeling that the password in /etc/pki/vdsm/keys/libvirt_password is static for all installations. And gerrit proves me right: http://gerrit.ovirt.org/gitweb?p=vdsm.git;a=blob;f=vdsm/libvirt_password;h=09e60bce9bc401bb8943154f7cb9cb08bd0f49da;hb=refs/heads/master So what is the purpose of authentication when that information is public? I created a BZ for this: https://bugzilla.redhat.com/show_bug.cgi?id=1061639 PS: I hope, whoever coded this, feels a little bit ashamed and perhaps buys a good book on writing secure code and reads it.. Am 05.02.2014 09:55, schrieb Joop: On 5-2-2014 9:27, Sven Kieske wrote: Well, I just tested this. and I can't connect to virsh with this information. I guess the mentioned user vdsm@rhevh might not be the actual one used in 3.3.2 anymore? (mail is from 2012 and mentions rhev, so..) vdsm@ovirt seems to work :-) Joop -- 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] Will this two node concept scale and work?
Hello List, my aim is to host multiple VMs which are redundant and are high available. It should also scale well. I think usually people just buy a fat iSCSI Storage and attach this. In my case it should scale well from very small nodes to big ones. Therefore an iSCSI Target will bring a lot of overhead (10GBit Links and two Paths, and really i should have a 2nd Hot Standby SAN, too). This makes scalability very hard. This post is also not meant to be a iscsi discussion. Since oVirt does not support DRBD out of the box i came up with my own concept: http://oi62.tinypic.com/2550xg5.jpg As far as i can tell i have the following advantages: - i can start with two simple cheap nodes - i could add more disks to my nodes. Maybe even a SSD as a dedicated drbdresource. - i can connect the two nodes directly to each other with bonding or infiniband. i dont need a switch or something between it. Downside: --- - i always need two nodes (as a couple) Will this setup work for me. So far i think i will be quite happy with it. Since the DRBD Resources are shared in dual primary mode i am not sure if ovirt can handle it. It is not allowed to write to a vm disk at the same time. The Concept of Linbit ( http://www.linbit.com/en/company/news/333-high-available-virtualization-at-a-most-reasonable-price) seems to much of an overhead with the iSCSI Layer and pacemaker setup. Its just too much for such a simple task. Please tell me that this concept is great and will work and scale well. Otherwise i am also thankful for any hints or critical ideas. Thanks a lot, Mario ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Engine-setup Fail
Yes after the download with yum, engine-setup works without fail. I up 3.4.0 to 3.5.0. I can send you log files when engine-setup fail and when he works. I have just one ovirt-engine so i can't reproduce this. Kev. 2014-02-05 Yedidyah Bar David d...@redhat.com: That's actually bad news, not good ones, because now we have no clue what the actual problem was :-) Anyway, principally this is unsupported. Usually this is done only during engine-setup when the engine is down - in order to prevent the new engine access the old database. After doing this you should immediately run engine-setup. Did you? Was it successful? Any chance you'll be able to try and reproduce this (on some other system if needed)? For the record, yours is the first such report I heard of. I still can't imagine how this can be a bug of ovirt, but what do I know :-( Thanks! -- Didi -- *From: *Kevin Tibi kevint...@hotmail.com *To: *Alon Bar-Lev alo...@redhat.com *Cc: *users users@ovirt.org *Sent: *Wednesday, February 5, 2014 8:37:27 AM *Subject: *Re: [Users] Engine-setup Fail Works ! Thx Alon. 2014-02-04 Alon Bar-Lev alo...@redhat.com: Can you please try to use yum for that, I do not think it is an issue of ovirt... yum --disableplugin=versionlock update ovirt-engine - Original Message - From: Kevin Tibi kevint...@hotmail.com To: Yedidyah Bar David d...@redhat.com Cc: users users@ovirt.org Sent: Tuesday, February 4, 2014 5:25:42 PM Subject: Re: [Users] Engine-setup Fail It's work if i down myself. But when i put the file in cache of yum, engine-setup download and bug. Maybe i put the file in the wrong directory. 2014-02-04 Yedidyah Bar David d...@redhat.com : What happens if you try downloading manually? E.g. wget http://resources.ovirt.org/releases/nightly/rpm/Fedora/19/noarch/ovirt-engine-userportal-3.5.0-0.0.master.20140130171441.git9257b30.fc19.noarch.rpm -- Didi From: Kevin Tibi kevint...@hotmail.com To: Alon Bar-Lev alo...@redhat.com Cc: users users@ovirt.org Sent: Tuesday, February 4, 2014 3:51:18 PM Subject: Re: [Users] Engine-setup Fail Don't work :/ Same fail 2014-02-04 Alon Bar-Lev alo...@redhat.com : I really do not understand what's wrong, it is not within the resources.ovirt.org as far as I can see. Please try: # yum clean all # yum clean dbcache Then try again... - Original Message - From: Kevin Tibi kevint...@hotmail.com To: Alon Bar-Lev alo...@redhat.com Cc: users users@ovirt.org Sent: Tuesday, February 4, 2014 3:23:12 PM Subject: Re: [Users] Engine-setup Fail Thx Alon, I tried your solution but : [ ERROR ] Yum [u'Des erreurs ont \xe9t\xe9 rencontr\xe9e durant le t\xe9l\xe9chargement des paquets.', u'ovirt-engine-userportal-3.5.0-0.0.master.20140130171441.git9257b30.fc19.noarch: failure: noarch/ovirt-engine-userportal-3.5.0-0.0.master.20140130171441.git9257b30.fc19.noarch.rpm from ovirt-nightly: [Errno 256] No more mirrors to try.\nhttp:// resources.ovirt.org/releases/nightly/rpm/Fedora/19/noarch/ovirt-engine-userportal-3.5.0-0.0.master.20140130171441.git9257b30.fc19.noarch.rpm : [Errno 14] HTTP Error 416 - Requested Range Not Satisfiable'] [ ERROR ] Failed to execute stage 'Package installation': [u'Des erreurs ont \xe9t\xe9 rencontr\xe9e durant le t\xe9l\xe9chargement des paquets.', u'ovirt-engine-userportal-3.5.0-0.0.master.20140130171441.git9257b30.fc19.noarch: failure: noarch/ovirt-engine-userportal-3.5.0-0.0.master.20140130171441.git9257b30.fc19.noarch.rpm from ovirt-nightly: [Errno 256] No more mirrors to try.\nhttp:// resources.ovirt.org/releases/nightly/rpm/Fedora/19/noarch/ovirt-engine-userportal-3.5.0-0.0.master.20140130171441.git9257b30.fc19.noarch.rpm : [Errno 14] HTTP Error 416 - Requested Range Not Satisfiable'] 2014-02-04 Alon Bar-Lev alo...@redhat.com : HTTP Error 416 - Requested Range Not Satisfiable Please use resources.ovirt.org and not ovirt.ovirt.org at you repo configuration. - Original Message - From: Kevin Tibi kevint...@hotmail.com To: users users@ovirt.org Sent: Tuesday, February 4, 2014 3:03:01 PM Subject: [Users] Engine-setup Fail Hi all, i try to update my ovirt-engine. I use nightly repo. For moment, i am on 3.4. When i try to up 3.5 i have a fail when it download the package: [ INFO ] Yum Downloading: (7/7): ovirt-engine-webadmin-portal-3.5.0-0.0.master.20140130171441.git9257b30.fc19.noarch.rpm 48 M(47%) [ ERROR ] Yum [u'Des erreurs ont \xe9t\xe9 rencontr\xe9e durant le t\xe9l\xe9chargement des paquets.', u'ovirt-engine-dbscripts-3.5.0-0.0.master.20140130171441.git9257b30.fc19.noarch: failure:
Re: [Users] Will this two node concept scale and work?
see in line On 02/05/2014 10:45 AM, ml ml wrote: Hello List, my aim is to host multiple VMs which are redundant and are high available. It should also scale well. I'm assuming you are talking about HA cluster since redundancy vm and HA vms are a contradiction :) I think usually people just buy a fat iSCSI Storage and attach this. In my case it should scale well from very small nodes to big ones. Therefore an iSCSI Target will bring a lot of overhead (10GBit Links and two Paths, and really i should have a 2nd Hot Standby SAN, too). This makes scalability very hard. This post is also not meant to be a iscsi discussion. Since oVirt does not support DRBD out of the box i came up with my own concept: check out posix storage domain. If it supports gluster you might be able to use it for DRBD. http://oi62.tinypic.com/2550xg5.jpg As far as i can tell i have the following advantages: - i can start with two simple cheap nodes - i could add more disks to my nodes. Maybe even a SSD as a dedicated drbd resource. - i can connect the two nodes directly to each other with bonding or infiniband. i dont need a switch or something between it. Downside: --- - i always need two nodes (as a couple) Will this setup work for me. So far i think i will be quite happy with it. Since the DRBD Resources are shared in dual primary mode i am not sure if ovirt can handle it. It is not allowed to write to a vm disk at the same time. not true that you cannot write to the same vm disk at the same time - you have a shared disk option The Concept of Linbit (http://www.linbit.com/en/company/news/333-high-available-virtualization-at-a-most-reasonable-price) seems to much of an overhead with the iSCSI Layer and pacemaker setup. Its just too much for such a simple task. Please tell me that this concept is great and will work and scale well. Otherwise i am also thankful for any hints or critical ideas. Thanks a lot, Mario ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users -- Dafna Ron ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Networking. Hosted Setup. All in One Host. Hetzner
On Thu, Jan 16, 2014 at 11:51:25PM +, Peter Styk wrote: Greetings, I'm writing here as to share some of my findings about hosting with Hetzner. All in one setups on single remote host can be tricky. Provider mounted an extra /29 subnet to the main host but none is routed by default and host has to become router itself. At the same time single mistake in bridging configuration and lost access results in need for re-bootstrap. It's still tempting to try and with many trials I eventually got to see guests talking to the net. Scenario 1: Working. Package bridge-utils, oVirt engine, setup bridge, VDSM, add host to engine, add routing to host routing table. Networking by trial and error. Still something is not right. Occasionally on ping out I'm getting Redirect Host (New nexthop messages. http://styk.tv/wp-content/uploads/2014/01/oVirtHosted1_almost_working.png Unfortunately, I fail to understand what can be hampering your routing there. Assaf, do you have a guess? Which version of ovirt have you been using? Now, with source-routing implemented into ovirt-3.3, there is a danger in setting your own content into route-netname, as it would be overwritten if netname is reconfigured via Engine. Scenario 2: Dreaming. Private network with private router/dhcp/nat. Private 10.0.0.0/24 network. No problems with routing as gateway 10.0.0.1 would be on the same subnet. Thought of using pfSense but can't seem to bring up an instance with two network cards on two different networks. I thought this would be easy. Go to Networks, click create new network, type private, save ok. then go to new instance. point at iso, attach two network cards. save ok. Launch Host did not satisfy internal filter Network No idea what that is. Maybe I don't understand how this works. I do not understand where having an instance with two nics fail. Is the Host did not satisfy internal filter Network message coming from Hetzner management, or oVirt's? I even tried removing ovirtmgmt network and leaving private network by itself. Tried with all 3 network card types (rtl8139/e1000/VirtIO) http://styk.tv/wp-content/uploads/2014/01/oVirtHosted2_preferred.png Could you explain how you configured your provate network? In my experience, your easiest option is to define a dummy interface ip link add name dummy_private type dummy and set up a normal oVirt network on top of it, as if it were a true nic. Either way if you have a minute or two please take a look at both attached diagrams. Deliberately making it difficult by forcing all elements on single box in hosted environment. Maybe there is a way to have this all installed with Neutron or vSwitch on the same box or is that pushing it? ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Networking. Hosted Setup. All in One Host. Hetzner
Could you explain further why does the host need to do any routing? Assaf Muller, Cloud Networking Engineer Red Hat - Original Message - From: Dan Kenigsberg dan...@redhat.com To: Peter Styk polf...@gmail.com, amul...@redhat.com Cc: users@ovirt.org Sent: Wednesday, February 5, 2014 1:23:54 PM Subject: Re: [Users] Networking. Hosted Setup. All in One Host. Hetzner On Thu, Jan 16, 2014 at 11:51:25PM +, Peter Styk wrote: Greetings, I'm writing here as to share some of my findings about hosting with Hetzner. All in one setups on single remote host can be tricky. Provider mounted an extra /29 subnet to the main host but none is routed by default and host has to become router itself. At the same time single mistake in bridging configuration and lost access results in need for re-bootstrap. It's still tempting to try and with many trials I eventually got to see guests talking to the net. Scenario 1: Working. Package bridge-utils, oVirt engine, setup bridge, VDSM, add host to engine, add routing to host routing table. Networking by trial and error. Still something is not right. Occasionally on ping out I'm getting Redirect Host (New nexthop messages. http://styk.tv/wp-content/uploads/2014/01/oVirtHosted1_almost_working.png Unfortunately, I fail to understand what can be hampering your routing there. Assaf, do you have a guess? Which version of ovirt have you been using? Now, with source-routing implemented into ovirt-3.3, there is a danger in setting your own content into route-netname, as it would be overwritten if netname is reconfigured via Engine. Scenario 2: Dreaming. Private network with private router/dhcp/nat. Private 10.0.0.0/24 network. No problems with routing as gateway 10.0.0.1 would be on the same subnet. Thought of using pfSense but can't seem to bring up an instance with two network cards on two different networks. I thought this would be easy. Go to Networks, click create new network, type private, save ok. then go to new instance. point at iso, attach two network cards. save ok. Launch Host did not satisfy internal filter Network No idea what that is. Maybe I don't understand how this works. I do not understand where having an instance with two nics fail. Is the Host did not satisfy internal filter Network message coming from Hetzner management, or oVirt's? I even tried removing ovirtmgmt network and leaving private network by itself. Tried with all 3 network card types (rtl8139/e1000/VirtIO) http://styk.tv/wp-content/uploads/2014/01/oVirtHosted2_preferred.png Could you explain how you configured your provate network? In my experience, your easiest option is to define a dummy interface ip link add name dummy_private type dummy and set up a normal oVirt network on top of it, as if it were a true nic. Either way if you have a minute or two please take a look at both attached diagrams. Deliberately making it difficult by forcing all elements on single box in hosted environment. Maybe there is a way to have this all installed with Neutron or vSwitch on the same box or is that pushing it? ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Putting a Gluster host into an ISCSI DC
On 02/05/2014 04:48 PM, Gianluca Cecchi wrote: Hello, in a test infra I would to temporarily put a Host that is part of Gluster DC and has gluster volumes active into an ISCSI DC. I put Gluster DC in maintenance, stop glustervolume and put both hosts in maintenance. Then I try to edit host and put into the ISCSI DC. But I receive error Error while executing action: node02: Cannot Edit host. Server having Gluster volume As you said above, if you have volumes in the cluster, volumes needs to be deleted before moving the hosts to a different cluster. Is there any basic check so that I cannot do this at all? Any tip? Thanks, Gianluca ___ 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] NAT configuration
On Wed, Feb 05, 2014 at 09:50:04AM +, Sven Kieske wrote: I can confirm that vdsm@ovirt does work. However, I have the strong feeling that the password in /etc/pki/vdsm/keys/libvirt_password is static for all installations. And gerrit proves me right: http://gerrit.ovirt.org/gitweb?p=vdsm.git;a=blob;f=vdsm/libvirt_password;h=09e60bce9bc401bb8943154f7cb9cb08bd0f49da;hb=refs/heads/master So what is the purpose of authentication when that information is public? I created a BZ for this: https://bugzilla.redhat.com/show_bug.cgi?id=1061639 PS: I hope, whoever coded this, feels a little bit ashamed and perhaps buys a good book on writing secure code and reads it.. I feel ashamed, but not due to the security issue here. Vdsm uses a unix domain socket to connect to libvirtd. That socket is owned by vdsm, so that only vdsm and root can use it. There is no security reason to use a password at all. I am ashamed for caving in and adding an obfuscation layer, designed only to deter local administrators from messing with libvirt under the feet of ovirt. This little hurdle does not deter from messing with qemu directly, but I suppose that qemu's command line does a good job anyway. Red Hat support folks repeatedly claim that this hurdle is more effective than putting a release note warning of the dangers in direct libvirt access. Dan. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] NAT configuration
On Tue, Feb 04, 2014 at 08:20:26PM +, martin.tru...@cspq.gouv.qc.ca wrote: Hi, I want to configure NAT in Ovirt with this procedure : http://lists.ovirt.org/pipermail/users/2012-April/001751.html But I don't have the login password of virsh and for the PosgreSQL database. I used the last version of oVirt with default installation with engine-setup Note that now it is possible to use vdsm-hook-extnet instead of changing Engine's database (step 12 and forth). You do not have your NAT network defined in oVirt, but you can define a vNIC Profile with the property extnet=natbr0 and whatever network you have defined (say ovirtmgmt). When you attach a vnic of a VM to your vNIC profile and start the VM, the hook script kicks into action and points the vnic to natbr0. Dan. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Will this two node concept scale and work?
Hello Ron, thanks for your reply. This post is also not meant to be a iscsi discussion. Since oVirt does not support DRBD out of the box i came up with my own concept: check out posix storage domain. If it supports gluster you might be able to use it for DRBD. Sorry, I dont quite understand how posix will work with gluster here. What would the architechture look like and on what layer would it replicate. With gluster the glusterfs comes into my mind. How will drbd come into place here? Thanks a lot! Mario ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Networking. Hosted Setup. All in One Host. Hetzner
On Wed, Feb 05, 2014 at 06:36:32AM -0500, Assaf Muller wrote: Could you explain further why does the host need to do any routing? From what I gather, the hosting service (Hetzner) allows only IP traffic out of the box, but Peter may correct me. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Will this two node concept scale and work?
From: ml ml mliebher...@googlemail.com To: Users@ovirt.org Sent: Wednesday, February 5, 2014 12:45:55 PM Subject: [Users] Will this two node concept scale and work? Hello List, my aim is to host multiple VMs which are redundant and are high available. It should also scale well. I think usually people just buy a fat iSCSI Storage and attach this. In my case it should scale well from very small nodes to big ones. Therefore an iSCSI Target will bring a lot of overhead (10GBit Links and two Paths, and really i should have a 2nd Hot Standby SAN, too). This makes scalability very hard. This post is also not meant to be a iscsi discussion. Since oVirt does not support DRBD out of the box i came up with my own concept: http://oi62 . tinypic .com/2550xg5. jpg As far as i can tell i have the following advantages: - i can start with two simple cheap nodes - i could add more disks to my nodes. Maybe even a SSD as a dedicated drbd resource. - i can connect the two nodes directly to each other with bonding or infiniband . i dont need a switch or something between it. Downside: --- - i always need two nodes (as a couple) Will this setup work for me. So far i think i will be quite happy with it. Since the DRBD Resources are shared in dual primary mode i am not sure if ovirt can handle it. It is not allowed to write to a vm disk at the same time. I don't know ovirt enough to comment on that. I did play in the past with drbd and libvirt (virsh). Note that having both nodes primary all the time for all resources is calling for a disaster. In any case of split brain, for any reason, drbd will not know what to do. What I did was to allow both to be primary, but had only one primary most of the time (per resource). I wrote a script to do migration, which made both primary for the duration of the migration (required by qemu) and then moved the source to secondary when migration finished. This way you still have a chance for a disaster, if there is a problem (split brain, node failure) during a migration. So if you decide to go this way, carefully plan and test to see that it works well for you. One source for a split brain, for me, at the time, was buggy nic drivers and bad bonding configuration. So test that well too if applicable. The approach I took seems similar to DRBD on LV level in [1], but with custom scripts and without ovirt. You might be able to make ovirt do this for you with hooks. Didn't try that. An obvious downside to this approach is that if one node in a pair is down, the other has no backup now. If you have multiple nodes and external shared storage, multiple nodes can be down with no disruption to service if the remaining nodes are capable enough. [1] http://www.ovirt.org/Features/DRBD Best regards, -- Didi ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Will this two node concept scale and work?
On Wed, 2014-02-05 at 08:49 -0500, Yedidyah Bar David wrote: From: ml ml mliebher...@googlemail.com To: Users@ovirt.org Sent: Wednesday, February 5, 2014 12:45:55 PM Subject: [Users] Will this two node concept scale and work? Hello List, my aim is to host multiple VMs which are redundant and are high available. It should also scale well. I think usually people just buy a fat iSCSI Storage and attach this. In my case it should scale well from very small nodes to big ones. Therefore an iSCSI Target will bring a lot of overhead (10GBit Links and two Paths, and really i should have a 2nd Hot Standby SAN, too). This makes scalability very hard. This post is also not meant to be a iscsi discussion. Since oVirt does not support DRBD out of the box i came up with my own concept: http://oi62.tinypic.com/2550xg5.jpg As far as i can tell i have the following advantages: - i can start with two simple cheap nodes - i could add more disks to my nodes. Maybe even a SSD as a dedicated drbd resource. - i can connect the two nodes directly to each other with bonding or infiniband. i dont need a switch or something between it. Downside: --- - i always need two nodes (as a couple) Will this setup work for me. So far i think i will be quite happy with it. Since the DRBD Resources are shared in dual primary mode i am not sure if ovirt can handle it. It is not allowed to write to a vm disk at the same time. I don't know ovirt enough to comment on that. I did play in the past with drbd and libvirt (virsh). Note that having both nodes primary all the time for all resources is calling for a disaster. In any case of split brain, for any reason, drbd will not know what to do. I second that, had many problems without proper fencing and even with fencing. What I did was to allow both to be primary, but had only one primary most of the time (per resource). I wrote a script to do migration, which made both primary for the duration of the migration (required by qemu) and then moved the source to secondary when migration finished. This way you still have a chance for a disaster, if there is a problem (split brain, node failure) during a migration. So if you decide to go this way, carefully plan and test to see that it works well for you. One source for a split brain, for me, at the time, was buggy nic drivers and bad bonding configuration. So test that well too if applicable. The approach I took seems similar to DRBD on LV level in [1], but with custom scripts and without ovirt. You might be able to make ovirt do this for you with hooks. Didn't try that.\ You could use drbd9 but I haven't tested it extensively yet. DRBD 9 has primary on write so you have both sides on passive until one of the nodes want's to write. It should automatically become primary then. This has been done by linbit to decrease split brain and expand to more than two nodes. http://www.drbd.org/users-guide-9.0/s-automatic-promotion.html But I don't know why it shouldn't work, maybe not with the node image but you can make a node of a normal rhel/centos/fedora install. One problem I always have with drbd and RHEL/Centos is that when you don't pay for the Linbit support, you don't get access to the repo and drbd is an additional option on RHEL. On Centos and Fedora the version is always lagging behind, so I have to compile the kernel module everytime for a new version or kernel update. An obvious downside to this approach is that if one node in a pair is down, the other has no backup now. If you have multiple nodes and external shared storage, multiple nodes can be down with no disruption to service if the remaining nodes are capable enough. [1] http://www.ovirt.org/Features/DRBD Best regards, -- Didi ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] My first wiki page
Thank you, I'll update the wiky in a few days to include this info (and I'll use it myself). Regards, El 05/02/14 04:43, Sahina Bose escribió: On 02/03/2014 07:18 PM, Juan Pablo Lorier wrote: Hi, I've created my first wiki page and I'd like someone to review it and tell me if there's something that need to be changed (besides it does not have any style yet) The URL is http://www.ovirt.org/oVirt_Wiki:How_to_change_Gluster%27s_network_interface Regards, Firstly, thanks for putting this information up! Some comments - 1. when you use different IP addresses for engine -to -gluster host (say IP1) and gluster -to -gluster communication (say IP2), operations from ovirt engine like add brick or remove brick would fail (as brick is tried to be added with IP1 which gluster does not understand) To work around this, it is better to use a FQDN both for registering the host with engine and also to peer probe the host from gluster CLI. You could have multiple IP addresses on the host resolve to the same FQDN. 2. To reuse a brick directory, gluster provides the force option during volume creation as well as adding bricks. This is available from gluster 3.5 upwards. [Adding Vijay to correct me, if I'm wrong] ___ 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] Putting a Gluster host into an ISCSI DC
Are the volumes hosted on the host? On Feb 5, 2014 12:18 PM, Gianluca Cecchi gianluca.cec...@gmail.com wrote: Hello, in a test infra I would to temporarily put a Host that is part of Gluster DC and has gluster volumes active into an ISCSI DC. I put Gluster DC in maintenance, stop glustervolume and put both hosts in maintenance. Then I try to edit host and put into the ISCSI DC. But I receive error Error while executing action: node02: Cannot Edit host. Server having Gluster volume Is there any basic check so that I cannot do this at all? Any tip? Thanks, Gianluca ___ 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] Putting a Gluster host into an ISCSI DC
On Wed, Feb 5, 2014 at 4:10 PM, Itamar Heim wrote: Are the volumes hosted on the host? Yes, but I have put down all resources related to it ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] NAT configuration
Well I didn't know the exact background for this code and I can understand it from a management perspective, but from a sysadmin perspective it is useless (it does not prevent anything against an informed attacker) and may be even lead to false security assumptions (nobody can mess with libvirt, it's all authenticated). But thanks for pointing out the reasoning behind this, I still don't like it, but I can understand it. (Funny side fact: the very first thing we did, when we found that libvirt just allows authenticated access was to find out how to create our own user, and every admin asks at first: how can I access libvirt, when something goes wrong?) Am 05.02.2014 13:45, schrieb Dan Kenigsberg: On Wed, Feb 05, 2014 at 09:50:04AM +, Sven Kieske wrote: I can confirm that vdsm@ovirt does work. However, I have the strong feeling that the password in /etc/pki/vdsm/keys/libvirt_password is static for all installations. And gerrit proves me right: http://gerrit.ovirt.org/gitweb?p=vdsm.git;a=blob;f=vdsm/libvirt_password;h=09e60bce9bc401bb8943154f7cb9cb08bd0f49da;hb=refs/heads/master So what is the purpose of authentication when that information is public? I created a BZ for this: https://bugzilla.redhat.com/show_bug.cgi?id=1061639 PS: I hope, whoever coded this, feels a little bit ashamed and perhaps buys a good book on writing secure code and reads it.. I feel ashamed, but not due to the security issue here. Vdsm uses a unix domain socket to connect to libvirtd. That socket is owned by vdsm, so that only vdsm and root can use it. There is no security reason to use a password at all. I am ashamed for caving in and adding an obfuscation layer, designed only to deter local administrators from messing with libvirt under the feet of ovirt. This little hurdle does not deter from messing with qemu directly, but I suppose that qemu's command line does a good job anyway. Red Hat support folks repeatedly claim that this hurdle is more effective than putting a release note warning of the dangers in direct libvirt access. Dan. -- 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] New iSCSI domain problem when first input wrong password
Hello, I have a fedora 19 engine based on 3.3 final. I installed a fedora 19 host (configured as infrastructure server) and then installed from web gui. Then yum update of the host updated vdsm (see separate thread I'm going to open for this). I create iSCSI DC and add that host to it. When I try to create the new iSCSI storage domain (sw iscsi target on a CentOS 6.5 server) I initially input the wrong chap password so that discovery goes ok, while authentication fails. Then I put correct pwd but after apparently going on (I see the exposed lun) then gives error in webadmin gui I have 2014-Feb-05, 16:57 Failed to attach Storage Domain OV01 to Data Center ISCSI. (User: admin@internal) 2014-Feb-05, 16:57 Failed to attach Storage Domains to Data Center ISCSI. (User: admin@internal) 2014-Feb-05, 16:57 The error message for connection 192.168.230.101 iqn.2013-09.local.localdomain:c6iscsit.target11 (LUN 1IET_00010001) returned by VDSM was: Failed to setup iSCSI subsystem 2014-Feb-05, 16:57 Storage Domain OV01 was added by admin@internal A workaround to have it activated is then (see logs 18:11): - put host in maintenance - reactivate it - attach sd to DC - ok (but see watermark message I don't understand) 2014-Feb-05, 18:13 The system has reached the 80% watermark on the VG metadata area size on OV01. This is due to a high number of Vdisks or large Vdisks size allocated on this specific VG. 2014-Feb-05, 18:13 Storage Domains were attached to Data Center ISCSI by admin@internal 2014-Feb-05, 18:13 Storage Domain OV01 (Data Center ISCSI) was activated by admin@internal 2014-Feb-05, 18:13 Storage Pool Manager runs on Host ovnode03 (Address: 192.168.33.44). 2014-Feb-05, 18:12 Data Center is being initialized, please wait for initialization to complete. 2014-Feb-05, 18:12 State was set to Up for host ovnode03. 2014-Feb-05, 18:11 Host ovnode03 was activated by admin@internal. 2014-Feb-05, 18:11 Host ovnode03 was switched to Maintenance mode by admin@internal. See my vdsm.log and supervdsm.log here https://drive.google.com/file/d/0BwoPbcrMv8mval9XMnN4ZGoxa00/edit?usp=sharing https://drive.google.com/file/d/0BwoPbcrMv8mveC1KQmo1dFAtNzQ/edit?usp=sharing Possibly initially wrong password is not optimally managed when putting the correct one, as I see some traceback. Gianluca ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] ovirt 3.3.3 host deploy push an old vdsm
On 02/05/2014 06:40 PM, Gianluca Cecchi wrote: Hello, fedora 19 engine 3.3.3 with ovirt-host-deploy-1.1.3-1.fc19.noarch when I deploy a fedora 19 host from web admin gui I get on it vdsm-4.13.0-11.fc19.x86_64 (it seems updated at november 2013) If after that I explicitly enable ovirt 3.3.3 yum repo and run yum update on host it correctly passes to : vdsm-4.13.3-3.fc19.x86_64 Gianluca ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users engine will deploy the latest vdsm the host sees based on the repos configured on the host? ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] ovirt 3.3.3 host deploy push an old vdsm
- Original Message - From: Itamar Heim ih...@redhat.com To: Gianluca Cecchi gianluca.cec...@gmail.com, users users@ovirt.org Sent: Wednesday, February 5, 2014 8:17:38 PM Subject: Re: [Users] ovirt 3.3.3 host deploy push an old vdsm On 02/05/2014 06:40 PM, Gianluca Cecchi wrote: Hello, fedora 19 engine 3.3.3 with ovirt-host-deploy-1.1.3-1.fc19.noarch when I deploy a fedora 19 host from web admin gui I get on it vdsm-4.13.0-11.fc19.x86_64 (it seems updated at november 2013) If after that I explicitly enable ovirt 3.3.3 yum repo and run yum update on host it correctly passes to : vdsm-4.13.3-3.fc19.x86_64 Gianluca ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users engine will deploy the latest vdsm the host sees based on the repos configured on the host? indeed. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] ovirt 3.3.3 host deploy push an old vdsm
On Wed, Feb 5, 2014 at 7:19 PM, Alon Bar-Lev wrote: - Original Message - From: Itamar Heim engine will deploy the latest vdsm the host sees based on the repos configured on the host? indeed. I thought sort of archive of rpm packages embedded and delivered from engine itself... and I thought it was in ovirt-host-deploy package on engine I didn't have any ovirt repo configured on host at deploy stage. Why not use repo of engine? Gianluca ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] ovirt 3.3.3 host deploy push an old vdsm
- Original Message - From: Gianluca Cecchi gianluca.cec...@gmail.com To: Alon Bar-Lev alo...@redhat.com Cc: Itamar Heim ih...@redhat.com, users users@ovirt.org Sent: Wednesday, February 5, 2014 8:29:49 PM Subject: Re: [Users] ovirt 3.3.3 host deploy push an old vdsm On Wed, Feb 5, 2014 at 7:19 PM, Alon Bar-Lev wrote: - Original Message - From: Itamar Heim engine will deploy the latest vdsm the host sees based on the repos configured on the host? indeed. I thought sort of archive of rpm packages embedded and delivered from engine itself... and I thought it was in ovirt-host-deploy package on engine I didn't have any ovirt repo configured on host at deploy stage. Why not use repo of engine? Gianluca I also suggested this a while ago... However, the current implementation is to use the system repo. ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[Users] A/B network setup
I am having some issues wrapping my head around this but what I am trying to setup is a A/B testing environment with a 3node cluster. Each node has 2 nics, 1 for ovirtmgmt and 1 for vlaned A/B network. I guess what I am trying to understand is if ovirt is tagging the vlan's I setup and is properly passing that to the switch? This would allow me to have multiple subnets on a host that the switch also can see as multiple vlans. Right? I would still have to configure the switch to allow that port access to both vlans. -- Patrick Pierson ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Mixing tagged and untagged VLANs
I attempted to set ovirtmgmt to untagged and making it a VM network but that would not save. The configuration: ovirtmgmt: - Display Network - Migration Network - VM Network - NO VLAN ipmi: - VM Network - VLAN 2 When I go into a host's Network Interfaces and attempt to sync the removal of VLAN 1 from ovirtmgmt I get the following message: Cannot setup Networks. The following Network Interfaces can have only a single VM Logical Network, or at most one non-VM Logical Network and/or several VLAN Logical Networks: eth0. So simply having one untagged and the rest tagged doesn't work. Is that expected? The use case here is simply to minimize cabling and hardware requirements as our current hypervisors have two GbE interfaces and a single IB interface. Our connectivity requirements are to 3 networks, public (direct connected to switches we have no control over for campus) and private (internal LAN) and our dedicated IPMI / Out-of-Band network (internal LAN). To interface VMs with all 3 networks and only 2 uplinks we used a tagged VLAN on our primary private network switch to link up our IPMI network as long as interface is tagged VLAN 2. The private continues to work as-is by remaining untagged. At the moment since the untagged ovirtmgmt + tagged ipmi logical network isn't allowed in oVirt we've had to explicitly set VLAN 1 (private) to tagged for specific ports on our switch. While this isn't a huge issue, to us it's an unnecessary complication of something that should work. I'm able to achieve the desired untagged + tagged setup using bridges in Linux when configured by hand, so I know that it technically can be done. Thanks - Trey On Tue, Feb 4, 2014 at 2:11 AM, Assaf Muller amul...@redhat.com wrote: Is it not possible to have multiple untagged VLAN networks associated to one interface in oVirt? No, not at this time. You can have one untagged network and N tagged networks on the same device, but only up to one untagged network. If you need multiple untagged networks on a single device then you're very welcome to report an RFE :) Assaf Muller, Cloud Networking Engineer Red Hat - Original Message - From: Trey Dockendorf treyd...@gmail.com To: users users@ovirt.org Sent: Monday, February 3, 2014 10:45:44 PM Subject: [Users] Mixing tagged and untagged VLANs Using 3.3.2 I seem unable to mix tagged and untagged VLANs on a single interface. I'm trying to put the following logical networks on a host's eth0. ovirtmgmt: - Display Network - Migration Network - NOT VM Network - NO VLAN private: - VM network - NO VLAN ipmi: - VM Network - VLAN 2 In the host's network setup ovirtmgmt is already linked to eth0. If I attach 'ipmi (VLAN 2)' then try and attach 'private' the message is Cannot have more than one non-VLAN network on one interface. Same occurs if I try and attach 'private' when only 'ovirtmgmt' is assigned to eth0. Is it not possible to have multiple untagged VLAN networks associated to one interface in oVirt? Thanks - Trey ___ 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] Gluster question
I currently have a new setup running ovirt 3.3.3. I have a Gluster storage domain with roughly 2.5TB of usable space. Gluster is installed on the same systems as the ovirt hosts. The host break down is as follows Ovirt DC: 4 hosts in the cluster. Each host has 4 physical disks in a RAID 5. Each disk is 500GB. With the OS installed and configured I end up with 1.2TB of usable space left for my data volume Gluster volume: 4 bricks with 1.2TB of space per brick (Distribute Replicate leaves me with about 2.5TB in the storage domain) Does this setup give me enough fault tolerance to survive losing a host and have my HA vm automatically move to an available host and keep running?? ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Quotas
Thanks for the input. -Original Message- From: Gilad Chaplik [mailto:gchap...@redhat.com] Sent: Wednesday, February 05, 2014 3:48 AM To: Maurice James Cc: users@ovirt.org Subject: Re: [Users] Quotas Hi Maurice, With quota you can set a hard limit for the number of CPU consumed by the user, and by that limit the num of vms, other quota parameters (memory and storage) can be unlimited. Although these videos are for version 3.1 they are pretty useful, since there weren't any drastic change since then: http://www.youtube.com/watch?v=zazJ_fW05Qk http://www.youtube.com/watch?v=y8qUoRImimY you are more than welcome to send your progress/use-case/any other question :-) Thanks, Gilad. - Original Message - From: Maurice James midnightst...@msn.com To: users@ovirt.org Sent: Wednesday, February 5, 2014 3:47:51 AM Subject: [Users] Quotas How do I limit the number of VMs that a user can create using quotas? ___ 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] Import VMware
Ok. Thanks. Sounds do-able From: Ted Miller [mailto:tmil...@hcjb.org] Sent: Wednesday, February 05, 2014 3:44 PM To: Maurice James; users@ovirt.org Subject: RE: [Users] Import VMware From: Maurice James midnightst...@msn.com mailto:midnightst...@msn.com Do you know of the best way to get a vmware guest into ovirt without virt-v2v by chance? From: users-boun...@ovirt.org mailto:users-boun...@ovirt.org [mailto:users-boun...@ovirt.org] On Behalf Of Ted Miller On 2/4/2014 10:49 AM, Maurice James wrote: Is it possible to import vmware images into ovirt 3.3, Or is a running Esx instance still required? This bug https://rhn.redhat.com/errata/RHBA-2013-1749.html officially withdrew support for importing image files directly, because it didn't always work. Ted Miller _ From: Maurice James midnightst...@msn.com mailto:midnightst...@msn.com Do you know of the best way to get a vmware guest into ovirt without virt-v2v by chance? _ No. I had a gluster + sanlock problem take out my ovirt cluster (2 hosts), and I only have it partially back up. My dozen VMs are currently available only when my (dual boot) hardware isn't running oVirt. Or, to put it the other way, I can only run oVirt when I can take down the VMWare group, because I don't have spare hardware. Working on rebuilding one VM in KVM today (VMWare copy had a problem). The only way I have heard succeed is to use ESX/ESXi or the hollow pig method. Create a VM in ovirt, including the hard drive. Replace hard drive file with file from VMWare (or otherwise get data into file). Fiddle with VM hardware settings until it runs. Ted Miller Elkhart, IN ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [Users] Gluster question
There was another recent post about this but a sum up was: You must have power fencing to support VM HA otherwise they'll be an issue with the engine not knowing whether the VM is still running and not bring it up on a new host to avoid data corruption. Also make sure you have your quorum setup properly based on your replication scenario so you can withstand 1 host being lost. I don't believe they'll keep running in a sense because of the host being lost, but they would restart on another host. At least that's what I've noticed in my case. On Thu, Feb 6, 2014 at 1:04 PM, Maurice James midnightst...@msn.com wrote: I currently have a new setup running ovirt 3.3.3. I have a Gluster storage domain with roughly 2.5TB of usable space. Gluster is installed on the same systems as the ovirt hosts. The host break down is as follows Ovirt DC: 4 hosts in the cluster. Each host has 4 physical disks in a RAID 5. Each disk is 500GB. With the OS installed and configured I end up with 1.2TB of usable space left for my data volume Gluster volume: 4 bricks with 1.2TB of space per brick (Distribute Replicate leaves me with about 2.5TB in the storage domain) Does this setup give me enough fault tolerance to survive losing a host and have my HA vm automatically move to an available host and keep running?? ___ 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] Gluster question
Hmm. So in that case would I be able to drop the Gluster setup and use NFS each host and make sure power fencing is enabled? Will that still achieve fault tolerance, or is a replicated gluster still required? From: Andrew Lau [mailto:and...@andrewklau.com] Sent: Wednesday, February 05, 2014 9:17 PM To: Maurice James Cc: users Subject: Re: [Users] Gluster question There was another recent post about this but a sum up was: You must have power fencing to support VM HA otherwise they'll be an issue with the engine not knowing whether the VM is still running and not bring it up on a new host to avoid data corruption. Also make sure you have your quorum setup properly based on your replication scenario so you can withstand 1 host being lost. I don't believe they'll keep running in a sense because of the host being lost, but they would restart on another host. At least that's what I've noticed in my case. On Thu, Feb 6, 2014 at 1:04 PM, Maurice James midnightst...@msn.com mailto:midnightst...@msn.com wrote: I currently have a new setup running ovirt 3.3.3. I have a Gluster storage domain with roughly 2.5TB of usable space. Gluster is installed on the same systems as the ovirt hosts. The host break down is as follows Ovirt DC: 4 hosts in the cluster. Each host has 4 physical disks in a RAID 5. Each disk is 500GB. With the OS installed and configured I end up with 1.2TB of usable space left for my data volume Gluster volume: 4 bricks with 1.2TB of space per brick (Distribute Replicate leaves me with about 2.5TB in the storage domain) Does this setup give me enough fault tolerance to survive losing a host and have my HA vm automatically move to an available host and keep running?? ___ 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
Re: [Users] Gluster question
I'm not sure what you mean by NFS each host, but you'll need some way to at least ensure the data is available. Be that be replicated gluster or a centralized SAN etc. On Thu, Feb 6, 2014 at 1:21 PM, Maurice James midnightst...@msn.com wrote: Hmm. So in that case would I be able to drop the Gluster setup and use NFS each host and make sure power fencing is enabled? Will that still achieve fault tolerance, or is a replicated gluster still required? *From:* Andrew Lau [mailto:and...@andrewklau.com] *Sent:* Wednesday, February 05, 2014 9:17 PM *To:* Maurice James *Cc:* users *Subject:* Re: [Users] Gluster question There was another recent post about this but a sum up was: You must have power fencing to support VM HA otherwise they'll be an issue with the engine not knowing whether the VM is still running and not bring it up on a new host to avoid data corruption. Also make sure you have your quorum setup properly based on your replication scenario so you can withstand 1 host being lost. I don't believe they'll keep running in a sense because of the host being lost, but they would restart on another host. At least that's what I've noticed in my case. On Thu, Feb 6, 2014 at 1:04 PM, Maurice James midnightst...@msn.com wrote: I currently have a new setup running ovirt 3.3.3. I have a Gluster storage domain with roughly 2.5TB of usable space. Gluster is installed on the same systems as the ovirt hosts. The host break down is as follows Ovirt DC: 4 hosts in the cluster. Each host has 4 physical disks in a RAID 5. Each disk is 500GB. With the OS installed and configured I end up with 1.2TB of usable space left for my data volume Gluster volume: 4 bricks with 1.2TB of space per brick (Distribute Replicate leaves me with about 2.5TB in the storage domain) Does this setup give me enough fault tolerance to survive losing a host and have my HA vm automatically move to an available host and keep running?? ___ 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] Gluster question
OK I think I got it now. What I meant by NFS on each host was. In ovirt you can set up an NFS storage domain on each host and have the available to all hosts in the cluster From: Andrew Lau [mailto:and...@andrewklau.com] Sent: Wednesday, February 05, 2014 9:25 PM To: Maurice James Cc: users Subject: Re: [Users] Gluster question I'm not sure what you mean by NFS each host, but you'll need some way to at least ensure the data is available. Be that be replicated gluster or a centralized SAN etc. On Thu, Feb 6, 2014 at 1:21 PM, Maurice James midnightst...@msn.com mailto:midnightst...@msn.com wrote: Hmm. So in that case would I be able to drop the Gluster setup and use NFS each host and make sure power fencing is enabled? Will that still achieve fault tolerance, or is a replicated gluster still required? From: Andrew Lau [mailto:and...@andrewklau.com mailto:and...@andrewklau.com ] Sent: Wednesday, February 05, 2014 9:17 PM To: Maurice James Cc: users Subject: Re: [Users] Gluster question There was another recent post about this but a sum up was: You must have power fencing to support VM HA otherwise they'll be an issue with the engine not knowing whether the VM is still running and not bring it up on a new host to avoid data corruption. Also make sure you have your quorum setup properly based on your replication scenario so you can withstand 1 host being lost. I don't believe they'll keep running in a sense because of the host being lost, but they would restart on another host. At least that's what I've noticed in my case. On Thu, Feb 6, 2014 at 1:04 PM, Maurice James midnightst...@msn.com mailto:midnightst...@msn.com wrote: I currently have a new setup running ovirt 3.3.3. I have a Gluster storage domain with roughly 2.5TB of usable space. Gluster is installed on the same systems as the ovirt hosts. The host break down is as follows Ovirt DC: 4 hosts in the cluster. Each host has 4 physical disks in a RAID 5. Each disk is 500GB. With the OS installed and configured I end up with 1.2TB of usable space left for my data volume Gluster volume: 4 bricks with 1.2TB of space per brick (Distribute Replicate leaves me with about 2.5TB in the storage domain) Does this setup give me enough fault tolerance to survive losing a host and have my HA vm automatically move to an available host and keep running?? ___ 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
