Re: [pve-devel] qemu : add support for vlan trunks on net interfaces
Is there documentation on this feature ? On Fri, Jan 15, 2016 at 3:15 PM, Alexandre Derumierwrote: > This add support vlan trunks for qemu net interfaces. > > based on theses patches: > http://pve.proxmox.com/pipermail/pve-devel/2014-September/012730.html > > This works with OVS and linux vlan-aware bridge. > > details are in commits > > Regards, > > Alexandre > > ___ > pve-devel mailing list > pve-devel@pve.proxmox.com > http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] UUID for VM's and disks
Hi Dietmar, Using a pool per cluster will mean that you may need to split/shrink your placement group count to stay within Ceph best-practises. With using a fixed number of pools, you can calculate the correct number of PG's for your estimated OSD count. By using a UUID for the disks and having a shared ceph pool, we would be able to move VM's between the PVE clusters to balance around the Proxmox clustering limits without fear of overlapping disk ID's. Having UUID's for VM's is not as important, but just ensures that 3rd party applications are referencing a unique ID for a VM and removes the potential for basic coding errors removing a VM from the wrong cluster. On Mon, May 16, 2016 at 5:08 PM, Dietmar Maurerwrote: >> The use of VMID's for disks is becoming an issue for us with Ceph. We >> cannot have multiple PVE clusters sharing the same pool on Ceph as >> there will be duplicate VMID's across the clusters that will result in >> a clash of RBD names on Ceph. I am sure this would be an issue with >> iSCSI+ZFS as well. > > The suggestion is to use different pools. > > >> Using UUID's for VM's would have benefits when using an external >> application to manage multiple PVE clusters. > > Why? What application do you talk about exactly? > > ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] UUID for VM's and disks
Hi, Have the Promox VE team thought about using UUID's for VM's as well as for Disks ? The use of VMID's for disks is becoming an issue for us with Ceph. We cannot have multiple PVE clusters sharing the same pool on Ceph as there will be duplicate VMID's across the clusters that will result in a clash of RBD names on Ceph. I am sure this would be an issue with iSCSI+ZFS as well. Using UUID's for VM's would have benefits when using an external application to manage multiple PVE clusters. VMID's could be retained for ease of management when using the local PVE web management, but via the API and internally UUID's could be used. Regards, Andrew ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge
Thanks Alexandre. I will build something for testing this Friday. On 5/08/2015 5:24 pm, Alexandre DERUMIER aderum...@odiso.com wrote: BTW, can you test this method too ? eth0.10vmbrcustomer--(vlanX)--tapX auto vmbrcustomer1 iface vmbrcustomer1 inet manual bridge_vlan_aware yes bridge_ports eth0.10 bridge_stp off bridge_fd 0 pre-up ip link add link eth0 eth0.10 type vlan proto 802.1ad id 10 ? - Mail original - De: aderumier aderum...@odiso.com À: Andrew Thrift and...@networklabs.co.nz Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Mardi 4 Août 2015 14:02:46 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Another way, but I'm not sure it's working, is to tag 802.1ad on the physical interface eth0.10vmbrcustomer--(vlanX)--tapX auto vmbrcustomer1 iface vmbrcustomer1 inet manual bridge_vlan_aware yes bridge_ports eth0.10 bridge_stp off bridge_fd 0 pre-up ip link add link eth0 eth0.10 type vlan proto 802.1ad id 10 - Mail original - De: aderumier aderum...@odiso.com À: Andrew Thrift and...@networklabs.co.nz Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Mardi 4 Août 2015 12:22:47 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Hi Alexandre, Hi, We also use QinQ and have submitted patches for the previous network implementation that made use of a bridge in bridge design to achieve the QinQ functionality. They are also a new way to implement q-in-q with vlan aware bridge http://www.spinics.net/lists/linux-ethernet-bridging/msg05514.html ++ +---+p/u +--+ ++ +--+ |eth0|--|802.1ad|veth|802.1Q|--|vnet|--|VM| ++ |bridge | |bridge| ++ +--+ +---+ +--+ p/u: pvid/untagged Currently we have implemented 802.1Q bridge. for qinq, we need to create a root bridge, with 802.1ad enabled, linked through a veth pair to 802.1Q bridge. The qinq bridge is managed exactly in the same way than 802.1ad, but it's enabled with echo 0x88a8 /sys/class/net/XXX/bridge/vlan_protocol for example eth0vmbr0--(vlan10)---brigelink---bridgelinkpeervmbrcustomer--(vlanX)--tapX brctl addbr vmbr0 echo 0x88a8 /sys/class/net/vmbr0/bridge/vlan_protocol ip link add dev bridgelink type veth peer name bridgelinkpeer brctl addif vmbr0 bridgelink brctl addif vmbrcustomer1 bridgelinkpeer bridge vlan add dev bridgelink vid 10 pvid untagged something like that I can try to make a patch, but I don't have hardware which support q-in-q for testing. - Mail original - De: Andrew Thrift and...@networklabs.co.nz À: aderumier aderum...@odiso.com Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Mardi 4 Août 2015 10:49:26 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Hi Alexandre, We also use QinQ and have submitted patches for the previous network implementation that made use of a bridge in bridge design to achieve the QinQ functionality. The new vlan aware bridge implementation will be a lot cleaner. When your patches are ready we will test them and provide feedback. Thanks, Andrew On Tue, Jul 28, 2015 at 2:09 AM, Alexandre DERUMIER aderum...@odiso.com wrote: does somebody have tested my vlan bridges patches ? (note that that need iproute2 from debian sid, for vlan ranges) It's working really fine here, I'm looking to add a patch for Q-in-Q bridge too. (I think Stefan Priebe use them) - Mail original - De: aderumier aderum...@odiso.com À: Wolfgang Bumiller w.bumil...@proxmox.com Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Vendredi 24 Juillet 2015 18:49:18 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Why is `bridge_add_interface` now restricted to the firewall-else branch? I manage it like openvswitch, vlan tagging is always done on the main bridge, not firewall bridge. + if ($firewall) { + $create_firewall_bridge_linux($iface, $bridge, $tag); create_firewall_bridge_linux($iface, $bridge, $tag) have - $bridge_add_interface($bridge, $vethfwpeer); + $bridge_add_interface($bridge, $vethfwpeer, $tag); #tag on the main bridge - return $fwbr; + $bridge_add_interface($fwbr, $iface); # add vm tap interface on fwbridge without vlan tag - Mail original - De: Wolfgang Bumiller w.bumil...@proxmox.com À: aderumier aderum...@odiso.com Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Vendredi 24 Juillet 2015 15:20:06 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge On Fri, Jul 24, 2015 at 01:52:59PM +0200, Alexandre Derumier wrote: - $newbridge = $create_firewall_bridge_linux($iface, $newbridge) if $firewall; + if (!$vlan_aware) { + my $newbridge = activate_bridge_vlan($bridge, $tag); + copy_bridge_config($bridge, $newbridge
Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge
Hi Alexandre, We also use QinQ and have submitted patches for the previous network implementation that made use of a bridge in bridge design to achieve the QinQ functionality. The new vlan aware bridge implementation will be a lot cleaner. When your patches are ready we will test them and provide feedback. Thanks, Andrew On Tue, Jul 28, 2015 at 2:09 AM, Alexandre DERUMIER aderum...@odiso.com wrote: does somebody have tested my vlan bridges patches ? (note that that need iproute2 from debian sid, for vlan ranges) It's working really fine here, I'm looking to add a patch for Q-in-Q bridge too. (I think Stefan Priebe use them) - Mail original - De: aderumier aderum...@odiso.com À: Wolfgang Bumiller w.bumil...@proxmox.com Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Vendredi 24 Juillet 2015 18:49:18 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Why is `bridge_add_interface` now restricted to the firewall-else branch? I manage it like openvswitch, vlan tagging is always done on the main bridge, not firewall bridge. + if ($firewall) { + $create_firewall_bridge_linux($iface, $bridge, $tag); create_firewall_bridge_linux($iface, $bridge, $tag) have - $bridge_add_interface($bridge, $vethfwpeer); + $bridge_add_interface($bridge, $vethfwpeer, $tag); #tag on the main bridge - return $fwbr; + $bridge_add_interface($fwbr, $iface); # add vm tap interface on fwbridge without vlan tag - Mail original - De: Wolfgang Bumiller w.bumil...@proxmox.com À: aderumier aderum...@odiso.com Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Vendredi 24 Juillet 2015 15:20:06 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge On Fri, Jul 24, 2015 at 01:52:59PM +0200, Alexandre Derumier wrote: - $newbridge = $create_firewall_bridge_linux($iface, $newbridge) if $firewall; + if (!$vlan_aware) { + my $newbridge = activate_bridge_vlan($bridge, $tag); + copy_bridge_config($bridge, $newbridge) if $bridge ne $newbridge; + $tag = undef; + } + + if ($firewall) { + $create_firewall_bridge_linux($iface, $bridge, $tag); + } else { + $bridge_add_interface($bridge, $iface, $tag); + } - $bridge_add_interface($newbridge, $iface); Why is `bridge_add_interface` now restricted to the firewall-else branch? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge
Alexandre, Am I right in thinking that for 4x bridges each with a different SVID (101,102,103,104) that the config would look like this ? auto vmbrcustomer1 iface vmbrcustomer1 inet manual bridge_vlan_aware yes bridge_ports customer1lp bridge_stp off bridge_fd 0 pre-up ip link add dev customer1l type veth peer name customer1lp post-up ip link set customer1l up auto vmbrcustomer2 iface vmbrcustomer2 inet manual bridge_vlan_aware yes bridge_ports customer2lp bridge_stp off bridge_fd 0 pre-up ip link add dev customer2l type veth peer name customer2lp post-up ip link set customer2l up auto vmbrcustomer3 iface vmbrcustomer3 inet manual bridge_vlan_aware yes bridge_ports customer3lp bridge_stp off bridge_fd 0 pre-up ip link add dev customer3l type veth peer name customer3lp post-up ip link set customer3l up auto vmbrcustomer4 iface vmbrcustomer4 inet manual bridge_vlan_aware yes bridge_ports customer4lp bridge_stp off bridge_fd 0 pre-up ip link add dev customer4l type veth peer name customer4lp post-up ip link set customer4l up auto vmbr0 iface vmbr0 inet manual bridge_vlan_aware yes bridge_ports eth0 customer1l customer2l customer3l customer4l bridge_stp off bridge_fd 0 post-up echo 0x88a8 /sys/class/net/vmbr0/bridge/vlan_protocol post-up bridge vlan add dev customer1l vid 101 pvid untagged post-up bridge vlan add dev customer2l vid 102 pvid untagged post-up bridge vlan add dev customer3l vid 103 pvid untagged post-up bridge vlan add dev customer4l vid 104 pvid untagged On Wed, Aug 5, 2015 at 11:54 AM, Andrew Thrift and...@networklabs.co.nz wrote: Hi Alexandre, This looks like it should work. Something to be aware of, QinQ does not always have an outer tag with ethertype 0x88a8, it can also have a tag of 0x8100 or 0x9100 depending on the implementation. For example: 0x88a8--0x8100Outer-tag (SVID) of 0x88a8, Inner-tag (CVID) of 0x8100 0x8100--0x8100Outer-tag (SVID) of 0x8100, Inner-tag (CVID) of 0x8100 0x9100--0x8100Outer-tag (SVID) of 0x9100, Inner-tag (CVID) of 0x8100 We typically use stacked tags of 0x8100 so both the outer and inner tag are 0x8100, this seems to be the most compatible method and is supported by all vendors that I am aware of. On Wed, Aug 5, 2015 at 12:25 AM, Alexandre DERUMIER aderum...@odiso.com wrote: This seem to work. (I'm not sure about tcpdump result when vlan are stacked) auto vmbrcustomer1 iface vmbrcustomer1 inet manual bridge_vlan_aware yes bridge_ports customer1lp bridge_stp off bridge_fd 0 pre-up ip link add dev customer1l type veth peer name customer1lp post-up ip link set customer1l up auto vmbr0 iface vmbr0 inet manual bridge_vlan_aware yes bridge_ports eth0 customer1l bridge_stp off bridge_fd 0 post-up echo 0x88a8 /sys/class/net/vmbr0/bridge/vlan_protocol post-up bridge vlan add dev customer1l vid 10 pvid untagged - Mail original - De: aderumier aderum...@odiso.com À: Andrew Thrift and...@networklabs.co.nz Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Mardi 4 Août 2015 14:02:46 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Another way, but I'm not sure it's working, is to tag 802.1ad on the physical interface eth0.10vmbrcustomer--(vlanX)--tapX auto vmbrcustomer1 iface vmbrcustomer1 inet manual bridge_vlan_aware yes bridge_ports eth0.10 bridge_stp off bridge_fd 0 pre-up ip link add link eth0 eth0.10 type vlan proto 802.1ad id 10 - Mail original - De: aderumier aderum...@odiso.com À: Andrew Thrift and...@networklabs.co.nz Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Mardi 4 Août 2015 12:22:47 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Hi Alexandre, Hi, We also use QinQ and have submitted patches for the previous network implementation that made use of a bridge in bridge design to achieve the QinQ functionality. They are also a new way to implement q-in-q with vlan aware bridge http://www.spinics.net/lists/linux-ethernet-bridging/msg05514.html ++ +---+p/u +--+ ++ +--+ |eth0|--|802.1ad|veth|802.1Q|--|vnet|--|VM| ++ |bridge | |bridge| ++ +--+ +---+ +--+ p/u: pvid/untagged Currently we have implemented 802.1Q bridge. for qinq, we need to create a root bridge, with 802.1ad enabled, linked through a veth pair to 802.1Q bridge. The qinq bridge is managed exactly in the same way than 802.1ad, but it's enabled with echo 0x88a8 /sys/class/net/XXX/bridge/vlan_protocol for example eth0vmbr0--(vlan10)---brigelink---bridgelinkpeer
Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge
Hi Alexandre, This looks like it should work. Something to be aware of, QinQ does not always have an outer tag with ethertype 0x88a8, it can also have a tag of 0x8100 or 0x9100 depending on the implementation. For example: 0x88a8--0x8100Outer-tag (SVID) of 0x88a8, Inner-tag (CVID) of 0x8100 0x8100--0x8100Outer-tag (SVID) of 0x8100, Inner-tag (CVID) of 0x8100 0x9100--0x8100Outer-tag (SVID) of 0x9100, Inner-tag (CVID) of 0x8100 We typically use stacked tags of 0x8100 so both the outer and inner tag are 0x8100, this seems to be the most compatible method and is supported by all vendors that I am aware of. On Wed, Aug 5, 2015 at 12:25 AM, Alexandre DERUMIER aderum...@odiso.com wrote: This seem to work. (I'm not sure about tcpdump result when vlan are stacked) auto vmbrcustomer1 iface vmbrcustomer1 inet manual bridge_vlan_aware yes bridge_ports customer1lp bridge_stp off bridge_fd 0 pre-up ip link add dev customer1l type veth peer name customer1lp post-up ip link set customer1l up auto vmbr0 iface vmbr0 inet manual bridge_vlan_aware yes bridge_ports eth0 customer1l bridge_stp off bridge_fd 0 post-up echo 0x88a8 /sys/class/net/vmbr0/bridge/vlan_protocol post-up bridge vlan add dev customer1l vid 10 pvid untagged - Mail original - De: aderumier aderum...@odiso.com À: Andrew Thrift and...@networklabs.co.nz Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Mardi 4 Août 2015 14:02:46 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Another way, but I'm not sure it's working, is to tag 802.1ad on the physical interface eth0.10vmbrcustomer--(vlanX)--tapX auto vmbrcustomer1 iface vmbrcustomer1 inet manual bridge_vlan_aware yes bridge_ports eth0.10 bridge_stp off bridge_fd 0 pre-up ip link add link eth0 eth0.10 type vlan proto 802.1ad id 10 - Mail original - De: aderumier aderum...@odiso.com À: Andrew Thrift and...@networklabs.co.nz Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Mardi 4 Août 2015 12:22:47 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Hi Alexandre, Hi, We also use QinQ and have submitted patches for the previous network implementation that made use of a bridge in bridge design to achieve the QinQ functionality. They are also a new way to implement q-in-q with vlan aware bridge http://www.spinics.net/lists/linux-ethernet-bridging/msg05514.html ++ +---+p/u +--+ ++ +--+ |eth0|--|802.1ad|veth|802.1Q|--|vnet|--|VM| ++ |bridge | |bridge| ++ +--+ +---+ +--+ p/u: pvid/untagged Currently we have implemented 802.1Q bridge. for qinq, we need to create a root bridge, with 802.1ad enabled, linked through a veth pair to 802.1Q bridge. The qinq bridge is managed exactly in the same way than 802.1ad, but it's enabled with echo 0x88a8 /sys/class/net/XXX/bridge/vlan_protocol for example eth0vmbr0--(vlan10)---brigelink---bridgelinkpeervmbrcustomer--(vlanX)--tapX brctl addbr vmbr0 echo 0x88a8 /sys/class/net/vmbr0/bridge/vlan_protocol ip link add dev bridgelink type veth peer name bridgelinkpeer brctl addif vmbr0 bridgelink brctl addif vmbrcustomer1 bridgelinkpeer bridge vlan add dev bridgelink vid 10 pvid untagged something like that I can try to make a patch, but I don't have hardware which support q-in-q for testing. - Mail original - De: Andrew Thrift and...@networklabs.co.nz À: aderumier aderum...@odiso.com Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Mardi 4 Août 2015 10:49:26 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Hi Alexandre, We also use QinQ and have submitted patches for the previous network implementation that made use of a bridge in bridge design to achieve the QinQ functionality. The new vlan aware bridge implementation will be a lot cleaner. When your patches are ready we will test them and provide feedback. Thanks, Andrew On Tue, Jul 28, 2015 at 2:09 AM, Alexandre DERUMIER aderum...@odiso.com wrote: does somebody have tested my vlan bridges patches ? (note that that need iproute2 from debian sid, for vlan ranges) It's working really fine here, I'm looking to add a patch for Q-in-Q bridge too. (I think Stefan Priebe use them) - Mail original - De: aderumier aderum...@odiso.com À: Wolfgang Bumiller w.bumil...@proxmox.com Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Vendredi 24 Juillet 2015 18:49:18 Objet: Re: [pve-devel] [PATCH] tap_plug : add support for vlan aware linux bridge Why is `bridge_add_interface` now restricted to the firewall-else branch? I manage it like openvswitch, vlan tagging is always done on the main bridge, not firewall bridge. + if ($firewall
[pve-devel] PVE-Enterprise ceph version
Hi Guys, There are a number of important fixes in the Ceph Firefly 0.80.9 release, in particular fixes for qemu+librbd performance. Is there any reason the PVE-Enterprise repository is still on 0.80.8 ? Can this be updated ? Thanks, ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] qemu-server: drive-mirror improvements
Thanks Alexandre, This sounds like it could fix the corruption issues we have been experiencing when doing drive-mirror's on RBD. On Tue, May 12, 2015 at 1:55 AM, Alexandre Derumier aderum...@odiso.com wrote: this is candidate for proxmox 3.4 fixes proxmox 4.0 First patch use new ready flag introduce in qemu 2.2, to finish block migration. It's the right way to known if we can complete the job. (same flag than qmp ready event) (I have also remove the vm pause when target is too slow, it don't work fine in my tests, and we can't rely on offset==len to known if the transfert is correctly finished. Second patch increse a lot a drive-mirror qmp, because the job is scanning the whole source image before return. This can be very long with big image on nfs for example. ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] add cloudinit support to proxmox ?
I think it is a brilliant idea. On Wed, May 6, 2015 at 5:24 AM, Alexandre DERUMIER aderum...@odiso.com wrote: Hi, I wonder if we could add some kind of cloudinit support to proxmox. It's already implemented in openstack,ovirt,rhev http://www.ovirt.org/Features/Cloud-Init_Integration and supported in almost all last guest distros (debian,centos, fedora, suse , and windows !) could be usefull to setup hostname,timezone, ip address to client. (we need to add some fields in vmid.conf, generate virtual iso or floppy for the first boot of the vm. What do you think about it ? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Current status of 3.10 kernels
We have been running 3.10 in production for over 2 years now. No real problems. On Mon, May 4, 2015 at 10:33 PM, Martin Waschbüsch serv...@waschbuesch.it wrote: Hi there, I have experienced some VM crashes when trying to run 64bit FreeBSD guests. Apparently, if I read the posts on the forum right, the 3.10 kernels do not have this problem. At any rate: is there a list of things known to work / not work with the latest 3.10 kernel? E.g. ceph, etc... Any caveats when upgrading a cluster, e.g. a certain order in which things should happen? Is it possible to upgrade one node and test things out without leaving it detached from the cluster? Thanks a lot for any and all input! Martin ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] qemu-server : balloning fix (finally !)
Did this patch make its way in to the pve-enterprise repository ? On Mon, Mar 9, 2015 at 8:22 PM, Alexandre Derumier aderum...@odiso.com wrote: The problem seem to come from polling interval, which is not setup automatically when -machine option exist. We already setup it manually, but not when we are doing migration. This patch always setup the pooling interval, if proxmox balloon option != 0 It's working fine with current proxmox pve-qemu-kvm, so no need to remove the balloon stats patch. ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] [PATCH] support QinQ / vlan stacking
So with the ebtables patch, bridge filter rules can be configured via the PVE API and WebUI ? On Fri, Mar 13, 2015 at 9:10 PM, Dietmar Maurer diet...@proxmox.com wrote: So it should be safe to use the vlan_filtering code? Maybe can we wait for rhel 7.1 kernel ? I think it should be pretty stable, but it need to be tested. Yes. I think that kernel should be available soon? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] balloon bug in qemu 2.1 ?
Hi Alexandre, Yes the affected VM seem to have -machine type=pc-i440fx-2.1 in their KVM parameters. I found a bunch that have machine: pc-i440fx-1.4 in the config, and KVM parameters, and they output the full balloon stats e.g.: # info balloon balloon: actual=4096 max_mem=4096 total_mem=4095 free_mem=2221 mem_swapped_in=0 mem_swapped_out=0 major_page_faults=0 minor_page_faults=43 last_update=1425629742 On Fri, Mar 6, 2015 at 4:17 AM, Alexandre DERUMIER aderum...@odiso.com wrote: ok,I speak too fast. It's not related to this patch. on current pve-qemu-kvm 2.1. balloon is working fine # info balloon balloon: actual=1024 max_mem=1024 total_mem=1002 free_mem=941 mem_swapped_in=0 mem_swapped_out=0 major_page_faults=120 minor_page_faults=215272 last_update=1425568324 But if the vm (qemu 2.1) is started with -machine type=pc-i440fx-2.1 or -machine type=pc-i440fx-2.0 (this is the case when you do a live migration) It's not working # info balloon balloon: actual=1024 max_mem=1024 @Andrew, are your vm where you see the info balloon bug, have been migrated from old proxmox (without stop/start ?) can you check in ssh if the kvm process have -machine type in the command line ? - Mail original - De: aderumier aderum...@odiso.com À: Andrew Thrift and...@networklabs.co.nz Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Jeudi 5 Mars 2015 16:05:03 Objet: Re: [pve-devel] balloon bug in qemu 2.1 ? I need to do more tests, but it seem that this commit (applied on qemu 2.2 but not on qemu 2.1) http://git.qemu.org/?p=qemu.git;a=commit;h=22644cd2c60151a964d9505f4c5f7baf845f20d8 fix the problem with qemu 2.1. (I have tested with the patch, balloon works fine, I need to test without the patch to compare) - Mail original - De: aderumier aderum...@odiso.com À: Andrew Thrift and...@networklabs.co.nz Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Jeudi 5 Mars 2015 15:41:51 Objet: Re: [pve-devel] balloon bug in qemu 2.1 ? in proxmox virtio-balloon-fix-query.patch, we have hw/virtio/virtio-balloon.c + + if (!(balloon_stats_enabled(dev) balloon_stats_supported(dev) + dev-stats_last_update)) { + return; + } + + info-last_update = dev-stats_last_update; + info-has_last_update = true; + + info-mem_swapped_in = dev-stats[VIRTIO_BALLOON_S_SWAP_IN]; + info-has_mem_swapped_in = info-mem_swapped_in = 0 ? true : false; + + info-mem_swapped_out = dev-stats[VIRTIO_BALLOON_S_SWAP_OUT]; + info-has_mem_swapped_out = info-mem_swapped_out = 0 ? true : false; + + info-major_page_faults = dev-stats[VIRTIO_BALLOON_S_MAJFLT]; + info-has_major_page_faults = info-major_page_faults = 0 ? true : false; + + info-minor_page_faults = dev-stats[VIRTIO_BALLOON_S_MINFLT]; + info-has_minor_page_faults = info-minor_page_faults = 0 ? true : false; + so, that mean that in qemu 2.1 + if (!(balloon_stats_enabled(dev) balloon_stats_supported(dev) + dev-stats_last_update)) { + return; + } one of this 3 funtions is not working - Mail original - De: Andrew Thrift and...@networklabs.co.nz À: aderumier aderum...@odiso.com Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Jeudi 5 Mars 2015 15:17:58 Objet: Re: [pve-devel] balloon bug in qemu 2.1 ? Hi Alexandre, This may be the cause of the crashes we have been experiencing. We reported them here: http://forum.proxmox.com/threads/21276-Kernel-Oops-Panic-on-3-10-5-and-3-10-7-Kernels These only started happening since we moved to qemu-2.1.x and we get the same output: # info balloon balloon: actual=16384 max_mem=16384 and have noticed VM's with only 1-2GB usage in the guest reporting almost the entire amount of ram used to the host, even though we have the latest balloon driver loaded and the blnsvr.exe service running. On Thu, Mar 5, 2015 at 11:33 PM, Alexandre DERUMIER aderum...@odiso.com wrote: Hi, I have see a bug report here: http://forum.proxmox.com/threads/2-RAM-Problem-since-Upgrade-to-3-4?p=108367posted=1#post108367 about balloon. on my qemu 2.2 #info balloon balloon: actual=1024 max_mem=2048 total_mem=985 free_mem=895 mem_swapped_in=0 mem_swapped_out=0 major_page_faults=301 minor_page_faults=61411 last_update=1425550707 same vm with qemu 2.2 + -machine type=pc-i440fx-2.1 #info balloon balloon: actual=1024 max_mem=2048 (Don't have true qemu 2.1 for test currently) ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel
Re: [pve-devel] balloon bug in qemu 2.1 ?
Hi Alexandre, This may be the cause of the crashes we have been experiencing.We reported them here: http://forum.proxmox.com/threads/21276-Kernel-Oops-Panic-on-3-10-5-and-3-10-7-Kernels These only started happening since we moved to qemu-2.1.x and we get the same output: # info balloon balloon: actual=16384 max_mem=16384 and have noticed VM's with only 1-2GB usage in the guest reporting almost the entire amount of ram used to the host, even though we have the latest balloon driver loaded and the blnsvr.exe service running. On Thu, Mar 5, 2015 at 11:33 PM, Alexandre DERUMIER aderum...@odiso.com wrote: Hi, I have see a bug report here: http://forum.proxmox.com/threads/2-RAM-Problem-since-Upgrade-to-3-4?p=108367posted=1#post108367 about balloon. on my qemu 2.2 #info balloon balloon: actual=1024 max_mem=2048 total_mem=985 free_mem=895 mem_swapped_in=0 mem_swapped_out=0 major_page_faults=301 minor_page_faults=61411 last_update=1425550707 same vm with qemu 2.2 + -machine type=pc-i440fx-2.1 #info balloon balloon: actual=1024 max_mem=2048 (Don't have true qemu 2.1 for test currently) ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] [PATCH] support QinQ / vlan stacking
We have found you need to nest the bridges to get QinQ to work in all scenarios. e.g. The above patch will work for MOST scenarios, but if you attach a vlan aware VM to the parent vmbr0 bridge it will cause traffic to the VM's to stop, or will not be able to see the tagged frames. The patch we use only has one other minor change: -activate_bridge_vlan_slave($bridgevlan, $iface, $tag); +activate_bridge_vlan_slave($bridgevlan, $bridge, $tag); On Sat, Feb 14, 2015 at 9:41 PM, Stefan Priebe s.pri...@profihost.ag wrote: Signed-off-by: Stefan Priebe s.pri...@profihost.ag --- data/PVE/Network.pm |2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/data/PVE/Network.pm b/data/PVE/Network.pm index 00639f6..97f4033 100644 --- a/data/PVE/Network.pm +++ b/data/PVE/Network.pm @@ -323,7 +323,7 @@ sub activate_bridge_vlan { my @ifaces = (); my $dir = /sys/class/net/$bridge/brif; -PVE::Tools::dir_glob_foreach($dir, '((eth|bond)\d+)', sub { +PVE::Tools::dir_glob_foreach($dir, '((eth|bond)\d+(\.\d+)?)', sub { push @ifaces, $_[0]; }); -- 1.7.10.4 ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Optimizing CPU flags
I agree. This would be of a big benefit to us, we quite often have to fix up VM's that users have created with sub-optimal disk and network configurations. On Wed, Feb 18, 2015 at 1:31 AM, Alexandre DERUMIER aderum...@odiso.com wrote: A fast workaround/hack would be to add such setting to datacenter.cfg I think also it could be great to be able to define default create values in datacenter.cfg (disk (virtio,scsi,...),nic (virtio,e1000,...), scsi controller (lsi,virtio-scsi), ...) - Mail original - De: dietmar diet...@proxmox.com À: Stefan Priebe s.pri...@profihost.ag, pve-devel pve-devel@pve.proxmox.com Envoyé: Mercredi 11 Février 2015 19:42:11 Objet: Re: [pve-devel] Optimizing CPU flags Sure but you cannot select a cpu type as a global default in a cluster. So you have to remember that one each time. The suggestion was to implement that: https://git.proxmox.com/?p=qemu-defaults.git;a=summary A fast workaround/hack would be to add such setting to datacenter.cfg ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] bridge on top of bridge?
Hi Stefan, We use bonded 10gigabit, our config is below. Note eth0 and eth1 are not used. We put customer VM's inside the relevant vmbr and tag them with a vlan. So say they are on vmbr0 and we set the vlan in PVE to 55 they would have an outer tag of 101 and an inner tag of 55. This works very nicely. auto lo iface lo inet loopback auto eth0 iface eth0 inet manual auto eth1 iface eth1 inet manual auto eth2 iface eth2 inet manual #post-up for i in rx tx sg tso gso gro; do ethtool -K $IFACE $i off; done auto eth3 iface eth3 inet manual #post-up for i in rx tx sg tso gso gro; do ethtool -K $IFACE $i off; done auto bond0 iface bond0 inet manual # bonding confuration bond_mode 802.3ad bond_miimon 100 bond-lacp-rate 1 bond-xmit-hash-policy layer3+4 bond_slaves eth2 eth3 post-up ifconfig $IFACE mtu 9000 #Cust01 vlan auto bond0.101 iface bond0.101 inet manual #Cust02 vlan auto bond0.102 iface bond0.102 inet manual #Cust03 vlan auto bond0.103 iface bond0.103 inet manual #Cust04 vlan auto bond0.104 iface bond0.104 inet manual #MgmtServers vlan auto bond0.200 iface bond0.200 inet manual #Bridge configuration auto mgmtbr0 iface mgmtbr0 inet static address 10.99.99.104 netmask 255.255.255.0 gateway 10.99.99.254 post-up ifconfig $IFACE mtu 1500 bridge_ports bond0.200 bridge_stp off bridge_fd 0 auto vmbr0 iface vmbr0 inet manual bridge_ports bond0.101 bridge_stp off bridge_fd 0 auto vmbr1 iface vmbr1 inet manual bridge_ports bond0.102 bridge_stp off bridge_fd 0 auto vmbr2 iface vmbr2 inet manual bridge_ports bond0.103 bridge_stp off bridge_fd 0 auto vmbr3 iface vmbr3 inet manual bridge_ports bond0.104 bridge_stp off bridge_fd 0 On Sat, Feb 14, 2015 at 9:11 PM, Stefan Priebe s.pri...@profihost.ag wrote: Hi, how does your interfaces file looks like? Thanks! Stefan Am 11.02.2015 um 22:28 schrieb Andrew Thrift: Hi Stefan, I posted a patch to do this a while back: http://pve.proxmox.com/pipermail/pve-devel/2013-April/006995.html It will need a few changes to work on pve-test but we have been running it in production for a few years now. On Thu, Feb 12, 2015 at 8:44 AM, Stefan Priebe s.pri...@profihost.ag mailto:s.pri...@profihost.ag wrote: Am 11.02.2015 um 20:25 schrieb Alexandre DERUMIER: does anybody know a way to have two bridges on top of one bond? AFAIK, it's not possible to put a single interfaces on 2 bridge. It's possible with tagged interfaces. Yes but the problem is i just would like to have a PVID - so no untagged traffic on that bridge but there may be other tap devices with a vid. But I think that mixing 1 interface - 1 brige 1 internface.tag - 1 another bridge is not working (maybe a linux bug). That may work but won't help me as the another bridge may contain multiple other tags. I think it's work with openvswitch. (I think it work also with an ovs on top of another ovs with tagged port) That sounds great. Never touched openvswitch? Is there a howto for proxmox? What's the advantage over normal linux bridges? Greets, Stefan - Mail original - De: Stefan Priebe s.pri...@profihost.ag mailto:s.pri...@profihost.ag À: pve-devel pve-devel@pve.proxmox.com mailto:pve-devel@pve.proxmox.com Envoyé: Mercredi 11 Février 2015 13:41:58 Objet: [pve-devel] bridge on top of bridge? Hi, does anybody know a way to have two bridges on top of one bond? I woulod like to use the PVID feature for another bridge than the default one (private LAN without the need to enable tagging on the tap device). I already tried to add the 2nd bridge on top of the bond or to put the 2nd bridge into the first bridge. Both does not work. Greets, Stefan _ pve-devel mailing list pve-devel@pve.proxmox.com mailto:pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-__bin/mailman/listinfo/pve-devel http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel _ pve-devel mailing list pve-devel@pve.proxmox.com mailto:pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-__bin/mailman/listinfo/pve-devel http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] [PATCH] iothreads : create 1 iothread by virtio device
Hi Alexandre, List Thanks for your very informative email. Our hosts are fairly powerful (E5-2697v2), as are our ceph hosts. We will run some more benchmarks shortly using KRBD + iothreads and report back. Regards, Andrew On Thu, Jan 29, 2015 at 9:20 PM, Alexandre DERUMIER aderum...@odiso.com wrote: Would be interesting to additionally benchmark with sheepdog, so that we can compare ;-) Why not, I'll try to see if I can do it. Also compare with a local raw file on ssd. - Mail original - De: dietmar diet...@proxmox.com À: aderumier aderum...@odiso.com Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Jeudi 29 Janvier 2015 08:54:43 Objet: Re: [pve-devel] [PATCH] iothreads : create 1 iothread by virtio device I going to have a full ssd ceph cluster next month (18x s 3500 on 3 hosts), with more powerfull nodes and more powerfull clients (20 cores 3,1Gghz) I'll do benchmarks and post results with differents setup. Would be interesting to additionally benchmark with sheepdog, so that we can compare ;-) ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Fwd: [Qemu-stable] [Qemu-devel] [PATCH v2 1/1] migration/block: fix pending() return value
I think you may be referring to a problem I reported. We have been having corruption of RBD volumes when live migrated from one pool to another inside Proxmox VE. Alexandre had spotted this bug and suggested it may be related. On Sun, Jan 4, 2015 at 8:37 AM, Stefan Priebe s.pri...@profihost.ag wrote: Am 03.01.2015 um 11:00 schrieb Dietmar Maurer: On January 2, 2015 at 5:37 PM Stefan Priebe - Profihost AG s.pri...@profihost.ag wrote: Isn't this something which was reported some weeks ago? Sorry, but what do you refer to? Sorry i thought there were a report where block migration was hanging. Stefan ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] [PATCH] rbd : free_image : retry if rbd has watchers
Hi Alexandre, yes this was tested with 3.3-3 # aptitude -F '%p %v#' search qemu-server qemu-server 3.3-3 On Mon, Dec 8, 2014 at 5:07 PM, Alexandre DERUMIER aderum...@odiso.com wrote: We are still experiencing corruption when performing live disk moves, even with this patch in place. Do you talk about patch rbd : free_image : retry if rbd has watchers ? because we don't have applied this patch, because it was not the problem but the consequence. Do you have tested last qemu-server package from pvetest repository ?(= 3.3-3) It should be fixed here. - Mail original - De: Andrew Thrift and...@networklabs.co.nz À: aderumier aderum...@odiso.com Cc: pve-devel pve-devel@pve.proxmox.com Envoyé: Lundi 8 Décembre 2014 04:42:23 Objet: Re: [pve-devel] [PATCH] rbd : free_image : retry if rbd has watchers Hi Alexandre, We are still experiencing corruption when performing live disk moves, even with this patch in place. Is there anything we can do to help pinpoint the cause of this ? On Wed, Nov 12, 2014 at 7:39 PM, Alexandre DERUMIER aderum...@odiso.com wrote: Ok, Great! Thanks for testing. (@cc pve-devel) - Mail transféré - De: Andrew Thrift and...@networklabs.co.nz À: Alexandre DERUMIER aderum...@odiso.com Envoyé: Mercredi 12 Novembre 2014 05:10:40 Objet: Re: [pve-devel] [PATCH] rbd : free_image : retry if rbd has watchers Hi Alexandre, Initial testing looks promising. I have tested migrating disks that have active writes on them and it worked well. All files had matching md5 sums. I will test with 4K writes tomorrow. On Fri, Nov 7, 2014 at 10:42 PM, Andrew Thrift and...@networklabs.co.nz wrote: Thanks Alexandre, I will try these first thing Monday. Have a good weekend ! On Fri, Nov 7, 2014 at 10:29 PM, Alexandre DERUMIER aderum...@odiso.com wrote: blockquote Do you know why online Disk Move's could be causing this corruption ? We have had to stop using it as if we corrupt a customers DB server it would not be a good thing :( Can you try to 2 patchs I have sent ? I think it should fix the problem. - Mail original - De: Andrew Thrift and...@networklabs.co.nz À: Alexandre DERUMIER aderum...@odiso.com Envoyé: Jeudi 6 Novembre 2014 21:18:19 Objet: Re: [pve-devel] [PATCH] rbd : free_image : retry if rbd has watchers HI Alexandre, Not related specifically to this patch. But using DIsk Move while the VM is online results in corruption for us almost every time we use it. We are using PVE3.3 with RBD storage. Typically we are moving from one RBD pool to another. We seem to get coorruption if the block copy completes or fails. We are primarily running Windows guest OS's with virtio or virtio-scsi disks. Our Ceph cluster has 84 spinning disks and 7x Intel S3700 Journal's. Networking to all devices is 2x10gigabit bonded and performance generally is very good. Do you know why online Disk Move's could be causing this corruption ? We have had to stop using it as if we corrupt a customers DB server it would not be a good thing :( On Fri, Nov 7, 2014 at 5:00 AM, Alexandre DERUMIER aderum...@odiso.com wrote: I'll resend a V2 tommorow - Mail original - De: Dietmar Maurer diet...@proxmox.com À: Alexandre DERUMIER aderum...@odiso.com Cc: pve-devel@pve.proxmox.com Envoyé: Jeudi 6 Novembre 2014 16:08:38 Objet: RE: [pve-devel] [PATCH] rbd : free_image : retry if rbd has watchers And what happens if we get other errors? Currently It's retrying until $i ~0 but we could add a die directly if $err !~ image still has watchers Yes, I think that would be better. ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel /blockquote ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] Remove Disk
Hi, I have noticed from at least PVE 2.3 through 3.3 that when removing a disk from a VM that there is no Task created, and that no progress is shown. This means that on storage such as Ceph where disk removal can take some time, you are in the dark as to progress. You also cannot see from a VM's Task History when a disk was deleted. Can this missing functionality please be added to the roadmap ? Thanks, Andrew ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Remove Disk
Hrmm I have had remove jobs run for 20+ minutes and never seen a task (ceph...) Wouldn't it be good practise to do this for all remove jobs ? I notice that when a Disk Move is performed that it shows the remove status. On Mon, Dec 8, 2014 at 2:40 PM, Alexandre DERUMIER aderum...@odiso.com wrote: Hi, I'm not sure, but I think a task is created only if the delete take too much time. I don't remember exactly after how much time. - Mail original - De: Andrew Thrift and...@networklabs.co.nz À: pve-devel pve-devel@pve.proxmox.com Envoyé: Lundi 8 Décembre 2014 00:12:50 Objet: [pve-devel] Remove Disk Hi, I have noticed from at least PVE 2.3 through 3.3 that when removing a disk from a VM that there is no Task created, and that no progress is shown. This means that on storage such as Ceph where disk removal can take some time, you are in the dark as to progress. You also cannot see from a VM's Task History when a disk was deleted. Can this missing functionality please be added to the roadmap ? Thanks, Andrew ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Maintenance mode
Option 4, You could have a drop down box with a default selection of Distribute Evenly then a list of all available nodes. This would allow the user to just click through and have them distribute evenly, OR they could select a specific node. On Tue, Nov 11, 2014 at 12:32 PM, Michael Rasmussen m...@datanom.net wrote: On Tue, 11 Nov 2014 00:27:46 +0100 Michael Rasmussen m...@datanom.net wrote: Hi all, Just to let you know. I have started implementing the feature for applying maintenance mode to a node. Right-click on any node will open a context menu which, at the moment, only contains one item: Maintenance. See attached screenshot. I quick poll. 1) Should the user select a specific target node? 2) Should the user select a list of target nodes? 3) Should the implementation migrate running VM's evenly across all other nodes? 4) Should the user be able to select any one of the methods above? -- Hilsen/Regards Michael Rasmussen Get my public GnuPG keys: michael at rasmussen dot cc http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xD3C9A00E mir at datanom dot net http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE501F51C mir at miras dot org http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE3E80917 -- /usr/games/fortune -es says: How should I know if it works? That's what beta testers are for. I only coded it. (Attributed to Linus Torvalds, somewhere in a posting) ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] [PATCH] rbd : free_image : retry if rbd has watchers
Hi Alexandre, Will this potentially fix the Disk Move issues I reported a few weeks back ? On Thu, Nov 6, 2014 at 4:09 AM, Alexandre Derumier aderum...@odiso.com wrote: reported in drive-mirror, if we cancel the job and we try to free the rbd volume, it seem that rbd storage take 2-3 seconds to release the volume or it's throwing an image still has watchers. This patch try to wait 10s with 10retries before dying Signed-off-by: Alexandre Derumier aderum...@odiso.com --- PVE/Storage/RBDPlugin.pm | 15 ++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/PVE/Storage/RBDPlugin.pm b/PVE/Storage/RBDPlugin.pm index 1026d81..3bfff4c 100644 --- a/PVE/Storage/RBDPlugin.pm +++ b/PVE/Storage/RBDPlugin.pm @@ -409,7 +409,20 @@ sub free_image { run_rbd_command($cmd, errmsg = rbd snap purge '$volname' error); $cmd = $rbd_cmd($scfg, $storeid, 'rm', $name); -run_rbd_command($cmd, errmsg = rbd rm '$volname' error); + +my $i = 0; +while(1){ + $i++; + eval { + run_rbd_command($cmd, errmsg = rbd rm '$volname' error); + }; + my $err = $@; + + sleep 1 if ($err $err =~ m/image still has watchers/); + + die image still has watchers if $i 10; + last if !$err; +} return undef; } -- 1.7.10.4 ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Guest UEFI support
Hi Dietmar, It is a replacement for the bios. You can load ovmf.fd with the -L option for QEMU. You can download binary copies of ovmf.fd from: http://tianocore.github.io/ovmf/ On Wed, Oct 29, 2014 at 12:22 AM, Dietmar Maurer diet...@proxmox.com wrote: I was wondering if support for UEFI in VM's could be added to PVE. See http://www.linux-kvm.org/page/OVMF It looks to be an alternative firmware image that is loaded instead of the usual bios.bin I would like to add EFI support, but those howtos are highly confusing for me. ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] Guest UEFI support
Hi, I was wondering if support for UEFI in VM's could be added to PVE. See http://www.linux-kvm.org/page/OVMF It looks to be an alternative firmware image that is loaded instead of the usual bios.bin Regards, Andrew ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] qemu-server : allow only hotpluggable|dynamics options to be change online
This seems like a very good idea. The web-ui could indicate the VM has pending changes, possibly by using a different VM icon, and/or showing it on the VM summary tab. Being able to view what changes are pending would also be a good idea in multi-admin environments. On Sat, Sep 6, 2014 at 2:11 AM, Dietmar Maurer diet...@proxmox.com wrote: Seems users also have different opinions - some prefer existing behavior Maybe simply add an option in datacenter.cfg, to choose behavior ? I think that is not necessary, because we already have the VM 'hotplug' option. I am thinking about the following: Write pending changes into a new config section, for example: --vm config--- memory: 1024 [pendign_changes] memory: 8192 -- We always write this before we change something. If hotplug is enabled, we commit the change after hotplug. Else, we commit on VM startup. What do you think? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] [PATCH] vm_devices_list : also list block devices
Some feedback. We have tested this patch and hot add/remove are both working perfectly with at least virtio-scsi on Windows 2008, 2008r2, 2012, 2012r2 and Linux 3.2 and higher. Thanks Guys! This is going to lower operational overheads for us as we wont need to schedule outages to cleanly add disks to VM's. On Mon, Sep 1, 2014 at 9:34 PM, Dietmar Maurer diet...@proxmox.com wrote: applied -Original Message- From: pve-devel [mailto:pve-devel-boun...@pve.proxmox.com] On Behalf Of Alexandre Derumier Sent: Freitag, 29. August 2014 15:04 To: pve-devel@pve.proxmox.com Subject: [pve-devel] [PATCH] vm_devices_list : also list block devices This allow scsi disk to be plug|unplug ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Pool View
Hi Dietmar, Sorry for the delay. Yes we have tested this on our test cluster and it appears to work as expected. Regards, Andrew On Wed, Aug 13, 2014 at 6:16 PM, Dietmar Maurer diet...@proxmox.com wrote: check_any call for valid_privs: This check is assessing whether the user has any permissions assigned to the pool and therefore, by extension, whether they should have visibility of the Pool at all, or whether the Pool should be omitted from their view of the list of Pools. In other words, if the user has no rights assigned to the Pool, then their console/screen should not be cluttered with it in this view (as they can not perform any actions on it). This check achieves that goal. Still don't get that. If a user has no permission assigned to a pool, the pool is not listed at all. If a user can see any VM inside that pool (VM.Audit), The view will show the pool. I cannot see any need for that check_any call. I just committed a first version of the Pool View: https://git.proxmox.com/?p=pve-manager.git;a=commitdiff;h=686915003e6069a25b9c8e3d533f770f0a71bebe Please can you test if that fits your needs? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Pool View
Exported Variable: The particular variable referenced is the canonical list of Valid Privileges, necessary for assessing whether any given Privilege is, in fact on the list of valid and assignable Privileges. The change made allows the variable to referenced outside the scope of the module that contains it; so that it can be re-used in other modules (specifically in PVE::API2::Cluster). I would like to highlight that it is NOT really a Global Variable. It is a lexically scoped, exported variable, available to other modules within the PVE package by reference to PVE:AccessControl::valid_privs (eg, it does not collide with other declarations of valid_privs in other modules). If a different approach is preferred, please advise what the desired strategy is - if it is strictly necessary, then for example a utility function could be added to PVE::AccessControl to export the values contained in the valid_privs variable instead. If the Proxmox has some specific design goals in mind the patch can be altered to accommodate. check_any call for valid_privs: This check is assessing whether the user has any permissions assigned to the pool and therefore, by extension, whether they should have visibility of the Pool at all, or whether the Pool should be omitted from their view of the list of Pools. In other words, if the user has no rights assigned to the Pool, then their console/screen should not be cluttered with it in this view (as they can not perform any actions on it). This check achieves that goal. !important tag on image url: This was included in the original cut of the patch to cover a specific browser incompatibility scenario; where the author observed that the image was not showing in its desired situation. This is a common problem with older versions of Internet Explorer and occasionally I have observed it in Chrome as well (where a less-specific CSS selector overrides a more-specific CSS selector). On Mon, Jul 21, 2014 at 5:40 PM, Dietmar Maurer diet...@proxmox.com wrote: here are a few comments and question (inline): diff -Nuarp usr/share/perl5/PVE/AccessControl.pm usr/share/perl5/PVE/AccessControl.pm --- usr/share/perl5/PVE/AccessControl.pm 2014-06-27 16:23:34.305587119 +1200 +++ usr/share/perl5/PVE/AccessControl.pm 2014-06-27 16:23:44.077692909 +1200 @@ -492,7 +492,7 @@ my $privgroups = { }, }; -my $valid_privs = {}; +our $valid_privs = {}; using global vars is bad design! my $special_roles = { 'NoAccess' = {}, # no priviledges diff -Nuarp usr/share/perl5/PVE/API2/Cluster.pm usr/share/perl5/PVE/API2/Cluster.pm --- usr/share/perl5/PVE/API2/Cluster.pm 2014-06-27 16:22:37.302970017 +1200 +++ usr/share/perl5/PVE/API2/Cluster.pm 2014-06-27 16:26:30.939493106 +1200 @@ -19,6 +19,8 @@ use PVE::RESTHandler; use PVE::RPCEnvironment; use PVE::JSONSchema qw(get_standard_option); +use PVE::AccessControl qw(valid_privs); + use base qw(PVE::RESTHandler); __PACKAGE__-register_method ({ @@ -160,12 +162,20 @@ __PACKAGE__-register_method({ my $vmlist = PVE::Cluster::get_vmlist() || {}; my $idlist = $vmlist-{ids} || {}; + my $storage_pool = {}; + + foreach my $pool (keys(%{$usercfg-{pools}})) { + foreach my $storage (keys(%{$usercfg-{pools}-{$pool}- {storage}})) { + $storage_pool-{$storage}=$pool; + } + } + my $pooldata = {}; if (!$param-{type} || $param-{type} eq 'pool') { foreach my $pool (keys %{$usercfg-{pools}}) { my $d = $usercfg-{pools}-{$pool}; - next if !$rpcenv-check($authuser, /pool/$pool, [ 'Pool.Allocate' ], 1); + next if !$rpcenv-check_any($authuser, /pool/$pool, [keys $PVE::AccessControl::valid_privs], 1); Please can you explain that change? my $entry = { id = /pool/$pool, @@ -185,9 +195,9 @@ __PACKAGE__-register_method({ my $data = $idlist-{$vmid}; my $entry = PVE::API2Tools::extract_vm_stats($vmid, $data, $rrd); - if ($entry-{uptime}) { - if (my $pool = $usercfg-{vms}-{$vmid}) { - if (my $pe = $pooldata-{$pool}) { + if (my $pool = $usercfg-{vms}-{$vmid}) { + if (my $pe = $pooldata-{$pool}) { + if ($entry-{uptime}) { $pe-{uptime} = $entry-{uptime} if !$pe- {uptime} || $entry-{uptime} $pe-{uptime}; $pe-{mem} = 0 if !$pe-{mem}; $pe-{mem} += $entry-{mem}; @@ -198,6 +208,7 @@ __PACKAGE__-register_method({ $pe-{maxcpu} = 0 if !$pe-{maxcpu}; $pe-{maxcpu} += $entry-{maxcpu}; } + $entry-{pool} = $pe-{pool}; } } @@ -228,6
Re: [pve-devel] KVM guest hangs with SCSI drive (ZFS)
This seems to be a trivial change, but fixes a known problem in PVE3.2 With such low risk, why would this NOT be included as an update to PVE3.2 ? On Tue, May 6, 2014 at 5:34 PM, Dietmar Maurer diet...@proxmox.com wrote: Will these patched packages (for qemu 1.7.1) make it in to the PVE3.2 repos ? No, my plan was to release them with qemu 2.0 ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] KVM guest hangs with SCSI drive (ZFS)
Dietmar, Is this not the point of pve-test ? If the updated packages are put in to pve-test users will be able to test this for you. We will test this against Ceph, iSCSI to EqualLogic, RAID via PERC5, iSCSI to Nexenta and NFS to Nexenta. On Tue, May 6, 2014 at 8:47 PM, Dietmar Maurer diet...@proxmox.com wrote: This seems to be a trivial change, but fixes a known problem in PVE3.2 With such low risk, why would this NOT be included as an update to PVE3.2 ? Updating a library is not a trivial change. This needs careful testing. ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] KVM guest hangs with SCSI drive (ZFS)
Will these patched packages (for qemu 1.7.1) make it in to the PVE3.2 repos ? On Tue, May 6, 2014 at 1:00 PM, Michael Rasmussen m...@datanom.net wrote: On Sun, 04 May 2014 18:47:28 +0200 (CEST) Alexandre DERUMIER aderum...@odiso.com wrote: can you test them ? Have tested with CentOS-6.5 and I can confirm that it works with virtio and megaraid. No speed monster though;-) -- Hilsen/Regards Michael Rasmussen Get my public GnuPG keys: michael at rasmussen dot cc http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xD3C9A00E mir at datanom dot net http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE501F51C mir at miras dot org http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE3E80917 -- /usr/games/fortune -es says: 1st graffitiist: QUESTION AUTHORITY! 2nd graffitiist: Why? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] PATCH: Add support for bridges with more than one physical link.
While this is a very neat way to load balance vlan traffic, it could be dangerous. You are effectively allowing users to create a loop. Unless they have their switches and spanning tree configured correctly upstream of the host, they could create a large broadcast storm on their network, likely knocking out other hosts and switches control planes. It is the same as looping a cable between two ports on a switch that does not have edge-safeguard functionality. Just my 2c. On Wed, Feb 12, 2014 at 6:28 PM, Pablo Ruiz pablo.r...@gmail.com wrote: Hi, In our proxmox cluster, each node has two bond interfaces, and each bond interface connects to and independent switch. This allows us to enable MSTP/PVSTP+ and thus load share traffic on different vlans across switches. +==+ | SWITCH-A |---, +==+ | +===+ | | | |-(bond1)--´ | -| Node-X | (trunk) | |-(bond2)--, | +===+ | | +==+ | | SWITCH-B |---´ +==+ In this setup, we have a couple of vlans (iSCSI-A iSCSI-B) each which has been priorized (by means of MSTP/PVST) on each switch. Also, proxmox's internal (software) bridges have STP disabled (so they do not conflict with MSTP's traffic). With this setup we are able to achieve full-redundant network interconnects, while at the same time using both links/bonds for iSCSI traffic (with multipath+round-robin). However, proxmox's current code doesnt allow bridges with more than one physical interface, something we had to apply an small enhacement to proxmox in order to setup our cluster as stated. We would like to have this enhacement merged into proxmox, and so I've read about proxmox development policies, etc. And as stated here is the link containing a diff format patch: https://github.com/pruiz/pve-common/commit/ce0173a1079e4fc8bb08d9eebd1df71f0f8dc3fe.diff aswell as the prettified diff from github: https://github.com/pruiz/pve-common/commit/ce0173a1079e4fc8bb08d9eebd1df71f0f8dc3fe This code has been in production for little more than a month with no issues. But, please let me know what maybe missing and/or what amendments needs to be done in order for this patch to be accepted into proxmox. Best regards, Pablo PD: I'll be sending the signed contribution aggrement by tomorrow, as soon as I get to my office. As I hope to send another contribution regarding ZFS plugin next. ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] PATCH: Add support for bridges with more than one physical link.
Correct in your case, but not all Proxmox users have MSTP/PVST+ configured. A much lower risk strategy is to use Multi-Chassis Link Aggregation Groups between Switch-A and Switch-B with a 4 port bond on the Proxmox hosts, using a L3/L4 hash you will achieve better load balancing of traffic across all 4 links and as this requires no changes to Proxmox will not introduce the chance of loops. We have had this in production for 13 months now with no issues. On Wed, Feb 12, 2014 at 9:55 PM, Pablo Ruiz pablo.r...@gmail.com wrote: That's what MSTP/PVSTP+ is supposed to avoid. (And infact, it does so in our environment).. however, it requires switches with such capability. On Wed, Feb 12, 2014 at 9:53 AM, Andrew Thrift and...@networklabs.co.nzwrote: While this is a very neat way to load balance vlan traffic, it could be dangerous. You are effectively allowing users to create a loop. Unless they have their switches and spanning tree configured correctly upstream of the host, they could create a large broadcast storm on their network, likely knocking out other hosts and switches control planes. It is the same as looping a cable between two ports on a switch that does not have edge-safeguard functionality. Just my 2c. On Wed, Feb 12, 2014 at 6:28 PM, Pablo Ruiz pablo.r...@gmail.com wrote: Hi, In our proxmox cluster, each node has two bond interfaces, and each bond interface connects to and independent switch. This allows us to enable MSTP/PVSTP+ and thus load share traffic on different vlans across switches. +==+ | SWITCH-A |---, +==+ | +===+ | | | |-(bond1)--´ | -| Node-X | (trunk) | |-(bond2)--, | +===+ | | +==+ | | SWITCH-B |---´ +==+ In this setup, we have a couple of vlans (iSCSI-A iSCSI-B) each which has been priorized (by means of MSTP/PVST) on each switch. Also, proxmox's internal (software) bridges have STP disabled (so they do not conflict with MSTP's traffic). With this setup we are able to achieve full-redundant network interconnects, while at the same time using both links/bonds for iSCSI traffic (with multipath+round-robin). However, proxmox's current code doesnt allow bridges with more than one physical interface, something we had to apply an small enhacement to proxmox in order to setup our cluster as stated. We would like to have this enhacement merged into proxmox, and so I've read about proxmox development policies, etc. And as stated here is the link containing a diff format patch: https://github.com/pruiz/pve-common/commit/ce0173a1079e4fc8bb08d9eebd1df71f0f8dc3fe.diff aswell as the prettified diff from github: https://github.com/pruiz/pve-common/commit/ce0173a1079e4fc8bb08d9eebd1df71f0f8dc3fe This code has been in production for little more than a month with no issues. But, please let me know what maybe missing and/or what amendments needs to be done in order for this patch to be accepted into proxmox. Best regards, Pablo PD: I'll be sending the signed contribution aggrement by tomorrow, as soon as I get to my office. As I hope to send another contribution regarding ZFS plugin next. ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface
Ok, so in my example of eth0,eth1,eth2,eth3bond0- ---bond0.101vmbr0vmbr0.201tap interface The packet is DOUBLE TAGGED with 0x8100 tags. So the outer tag (SVID) is 101 with an inner (CVID) tag of 201. 802.1ad support adds the capability to have an SVID of 0x88a8 with a CVID of 0x8100. Before this support was added you could only do QinQ using the stacked 0x8100 tags. On Mon, Jan 13, 2014 at 9:31 PM, Alexandre DERUMIER aderum...@odiso.comwrote: Also,I see that support of 802.1ad has been added to kernel since 3.10 only. So how do is work before ? http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=8ad227ff89a7e6f05d07cd0acfd95ed3a24450ca # ip link add link eth0 eth0.1000 \ type vlan proto 802.1ad id 1000 - Mail original - De: Alexandre DERUMIER aderum...@odiso.com À: Stefan Priebe - Profihost AG s.pri...@profihost.ag Cc: pve-devel@pve.proxmox.com Envoyé: Lundi 13 Janvier 2014 09:10:59 Objet: Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface this should explain it: http://en.wikipedia.org/wiki/IEEE_802.1ad Thanks, I understand now. So, for this setup: bond0.101vmbr0vmbr0.201tap interface when the packet from tap (tagged 201) is going out through bond0.101, what happen ? is the packet vlan retagged 101 ? or is the packet vlan 201 encapsuled in vlan101 ? - Mail original - De: Stefan Priebe - Profihost AG s.pri...@profihost.ag À: Alexandre DERUMIER aderum...@odiso.com, Andrew Thrift and...@networklabs.co.nz Cc: pve-devel@pve.proxmox.com Envoyé: Lundi 13 Janvier 2014 08:24:19 Objet: Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface Am 13.01.2014 07:54, schrieb Alexandre DERUMIER: QinQ vlan tagging. can somebody explain me how qinq works exactly ? (I'm reading cisco doc, but I'm not sure to understand how tagging is working exactly) this should explain it: http://en.wikipedia.org/wiki/IEEE_802.1ad - Mail original - De: Andrew Thrift and...@networklabs.co.nz À: pve-devel@pve.proxmox.com Envoyé: Lundi 13 Janvier 2014 01:33:24 Objet: Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface FYI we are using vlan tagging on bridges with Proxmox in production for over a year now, initially on 2.6.32 kernel and then on 3.10. We are using Intel gigabit and 10gigabit adapters. We posted the patches to the list a few months back, I believe these are very similar to Alexandre's patches. We have a more complex config in that we are also doing bonding and QinQ vlan tagging. Our setup looks like this: eth0,eth1,eth2,eth3bond0bond0.101vmbr0vmbr0.201tap interface That is using an outer tag of 101 and an inner tag of 201. On Sat, Jan 11, 2014 at 7:59 AM, Alexandre DERUMIER aderum...@odiso.com wrote: If alexandre’s patch don’t work with any devices it isn’t really interesting because it addressing other functionality and devices. I checked the patch and it use the same problematic part with eth*, wifi* and bond* check which fails with virtual devices like gre, ipsec,.. What do you mean by don't work with any devices ? My patch is to manage vlan tags on the bridge, not eth interface. eth0vmbr0--tap interface vlan are tagged on tap interfaces plugged on vmbr0, with new bridge cmd. (like an access port on a cisco switch) and vlans are allowed to pass through eth0.(like a trunk port on cisco switch) So I think it should work with gre,ipsec,...(But I don't have tested it yet) - Mail original - De: Johannes Ernst i...@filemedia.de À: pve-devel@pve.proxmox.com Envoyé: Vendredi 10 Janvier 2014 18:16:30 Objet: Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface Thus, it’s a configuration issue and NOT a kernel issue. If alexandre’s patch don’t work with any devices it isn’t really interesting because it addressing other functionality and devices. I checked the patch and it use the same problematic part with eth*, wifi* and bond* check which fails with virtual devices like gre, ipsec,.. Am 10.01.2014 um 17:18 schrieb Dietmar Maurer diet...@proxmox.com : Sure? Do you have additional information? I think it's not correct and it works! We tested that a few times (and failed), so nobody is keen to test that again. We currently try to use the new bridge VLAN features - that looks more promising (see patches from Alexandre). ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve
Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface
I should add that some switches will only accept QinQ using an SVID of 0x88a8 and a CVID of 0x8100. Which is why support for it was needed on Linux. A lot of switches will allow you to specify the SVID type, either globally, per port or per vlan. On Tue, Jan 14, 2014 at 10:42 AM, Andrew Thrift and...@networklabs.co.nzwrote: Ok, so in my example of eth0,eth1,eth2,eth3bond0- ---bond0.101vmbr0vmbr0.201tap interface The packet is DOUBLE TAGGED with 0x8100 tags. So the outer tag (SVID) is 101 with an inner (CVID) tag of 201. 802.1ad support adds the capability to have an SVID of 0x88a8 with a CVID of 0x8100. Before this support was added you could only do QinQ using the stacked 0x8100 tags. On Mon, Jan 13, 2014 at 9:31 PM, Alexandre DERUMIER aderum...@odiso.comwrote: Also,I see that support of 802.1ad has been added to kernel since 3.10 only. So how do is work before ? http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=8ad227ff89a7e6f05d07cd0acfd95ed3a24450ca # ip link add link eth0 eth0.1000 \ type vlan proto 802.1ad id 1000 - Mail original - De: Alexandre DERUMIER aderum...@odiso.com À: Stefan Priebe - Profihost AG s.pri...@profihost.ag Cc: pve-devel@pve.proxmox.com Envoyé: Lundi 13 Janvier 2014 09:10:59 Objet: Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface this should explain it: http://en.wikipedia.org/wiki/IEEE_802.1ad Thanks, I understand now. So, for this setup: bond0.101vmbr0vmbr0.201tap interface when the packet from tap (tagged 201) is going out through bond0.101, what happen ? is the packet vlan retagged 101 ? or is the packet vlan 201 encapsuled in vlan101 ? - Mail original - De: Stefan Priebe - Profihost AG s.pri...@profihost.ag À: Alexandre DERUMIER aderum...@odiso.com, Andrew Thrift and...@networklabs.co.nz Cc: pve-devel@pve.proxmox.com Envoyé: Lundi 13 Janvier 2014 08:24:19 Objet: Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface Am 13.01.2014 07:54, schrieb Alexandre DERUMIER: QinQ vlan tagging. can somebody explain me how qinq works exactly ? (I'm reading cisco doc, but I'm not sure to understand how tagging is working exactly) this should explain it: http://en.wikipedia.org/wiki/IEEE_802.1ad - Mail original - De: Andrew Thrift and...@networklabs.co.nz À: pve-devel@pve.proxmox.com Envoyé: Lundi 13 Janvier 2014 01:33:24 Objet: Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface FYI we are using vlan tagging on bridges with Proxmox in production for over a year now, initially on 2.6.32 kernel and then on 3.10. We are using Intel gigabit and 10gigabit adapters. We posted the patches to the list a few months back, I believe these are very similar to Alexandre's patches. We have a more complex config in that we are also doing bonding and QinQ vlan tagging. Our setup looks like this: eth0,eth1,eth2,eth3bond0bond0.101vmbr0vmbr0.201tap interface That is using an outer tag of 101 and an inner tag of 201. On Sat, Jan 11, 2014 at 7:59 AM, Alexandre DERUMIER aderum...@odiso.com wrote: If alexandre’s patch don’t work with any devices it isn’t really interesting because it addressing other functionality and devices. I checked the patch and it use the same problematic part with eth*, wifi* and bond* check which fails with virtual devices like gre, ipsec,.. What do you mean by don't work with any devices ? My patch is to manage vlan tags on the bridge, not eth interface. eth0vmbr0--tap interface vlan are tagged on tap interfaces plugged on vmbr0, with new bridge cmd. (like an access port on a cisco switch) and vlans are allowed to pass through eth0.(like a trunk port on cisco switch) So I think it should work with gre,ipsec,...(But I don't have tested it yet) - Mail original - De: Johannes Ernst i...@filemedia.de À: pve-devel@pve.proxmox.com Envoyé: Vendredi 10 Janvier 2014 18:16:30 Objet: Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface Thus, it’s a configuration issue and NOT a kernel issue. If alexandre’s patch don’t work with any devices it isn’t really interesting because it addressing other functionality and devices. I checked the patch and it use the same problematic part with eth*, wifi* and bond* check which fails with virtual devices like gre, ipsec,.. Am 10.01.2014 um 17:18 schrieb Dietmar Maurer diet...@proxmox.com : Sure? Do you have additional information? I think it's not correct and it works! We tested that a few times (and failed), so nobody is keen to test that again. We currently try to use the new bridge VLAN features - that looks more promising (see patches from Alexandre). ___ pve-devel mailing
Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface
FYI we are using vlan tagging on bridges with Proxmox in production for over a year now, initially on 2.6.32 kernel and then on 3.10. We are using Intel gigabit and 10gigabit adapters. We posted the patches to the list a few months back, I believe these are very similar to Alexandre's patches. We have a more complex config in that we are also doing bonding and QinQ vlan tagging. Our setup looks like this: eth0,eth1,eth2,eth3bond0bond0.101vmbr0vmbr0.201tap interface That is using an outer tag of 101 and an inner tag of 201. On Sat, Jan 11, 2014 at 7:59 AM, Alexandre DERUMIER aderum...@odiso.comwrote: If alexandre’s patch don’t work with any devices it isn’t really interesting because it addressing other functionality and devices. I checked the patch and it use the same problematic part with eth*, wifi* and bond* check which fails with virtual devices like gre, ipsec,.. What do you mean by don't work with any devices ? My patch is to manage vlan tags on the bridge, not eth interface. eth0vmbr0--tap interface vlan are tagged on tap interfaces plugged on vmbr0, with new bridge cmd. (like an access port on a cisco switch) and vlans are allowed to pass through eth0.(like a trunk port on cisco switch) So I think it should work with gre,ipsec,...(But I don't have tested it yet) - Mail original - De: Johannes Ernst i...@filemedia.de À: pve-devel@pve.proxmox.com Envoyé: Vendredi 10 Janvier 2014 18:16:30 Objet: Re: [pve-devel] [PATCH] Virtual vlan tagging to bridge interface Thus, it’s a configuration issue and NOT a kernel issue. If alexandre’s patch don’t work with any devices it isn’t really interesting because it addressing other functionality and devices. I checked the patch and it use the same problematic part with eth*, wifi* and bond* check which fails with virtual devices like gre, ipsec,.. Am 10.01.2014 um 17:18 schrieb Dietmar Maurer diet...@proxmox.com: Sure? Do you have additional information? I think it's not correct and it works! We tested that a few times (and failed), so nobody is keen to test that again. We currently try to use the new bridge VLAN features - that looks more promising (see patches from Alexandre). ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] Resource Pool problems when cloning
Hi, We have noticed on two separate clusters that when we perform a clone, and select a destination resource pool that the resulting cloned machine is NOT in the pool we selected (it is in the root pool) We also noticed that when logged in as a restricted user (that only has access to the one pool) that when we clone to the allowed resource pool it fails based on permissions. I am assuming that the second issue is related to the first, in that PVE is not actually putting the cloned VM in to the correct pool, it is putting it in to the default root pool. Is this a known issue ? If not, we are happy to create a clean patch for the issue and perform testing on it, if someone here is prepared to sponsor it. Regards, Andrew ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] Ops on multiple VM's
Hi, One of the features that was discussed for Proxmox 3.0 was the ability to perform an operation on multiple VM's at the same time. e.g. Select group of VM's or a Resource Group and Migrate them, or start/stop them. I was wondering if/when this feature will be implemented ? Regards, Andrew ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] Linked clones fail if source Ceph pool name is not rbd
Hi, On 3.0rc1 when I attempt to create a Linked-Clone from a VM on a pool other than rbd it fails with the following error: stopped: clone failed: rbd clone base-120-disk-1' error: rbd: error opening pool rbd: (2) No such file or directory Regards, Andrew ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] Cloned VM's do not clone parent cache settings
Good Morning, I have noticed that cloned VM's do not copy the cache settings of the Template. e.g. if I have a Template with write-back enabled, then perform a full clone it will have write-back cache disabled on the resulting VM. I am not sure if this is a feature, or a bug. Regards, Andrew ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] new ceph stable release !
Great! Will this fix make it in to the packages for 3.0 rc2 ? Regards, Andrew On 5/7/2013 10:21 PM, Dietmar Maurer wrote: Thanks for the hint: https://git.proxmox.com/?p=pve-qemu-kvm.git;a=commit;h=54619c4adf9014093ef09d0f28fd40e88f56a55a -Original Message- From: Stefan Priebe - Profihost AG [mailto:s.pri...@profihost.ag] Sent: Dienstag, 07. Mai 2013 11:51 To: Dietmar Maurer Cc: Alexandre DERUMIER; pve-devel@pve.proxmox.com Subject: Re: [pve-devel] new ceph stable release ! Keep in mind to update QEMU to fix the ugly rbd flush bug: http://git.qemu.org/?p=qemu.git;a=commitdiff;h=dc7588c1eb3008bda53dde1d 6b890cd299758155 ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Add sound card to VM's from WebUI
I have not been able to get sound over RDP working on Windows Server 2008r2 without a KVM hda sound card. As soon as this is added it works perfectly. I'll check our VMWare farms which always just work with RDP sound and see what we have there. On 4/19/2013 12:15 AM, Alexandre DERUMIER wrote: seem to be possible also for win2003 http://support.dvsweb.com/app/answers/detail/a_id/263/~/enabling-audio-over-rdp-on-a-windows-server,-without-hardware-soundcard - Mail original - De: Martin Maurer mar...@proxmox.com À: Andrew Thrift and...@networklabs.co.nz, Dietmar Maurer diet...@proxmox.com, pve-devel@pve.proxmox.com Envoyé: Jeudi 18 Avril 2013 13:49:39 Objet: Re: [pve-devel] Add sound card to VM's from WebUI Hi, -Original Message- From: pve-devel-boun...@pve.proxmox.com [mailto:pve-devel- boun...@pve.proxmox.com] On Behalf Of Andrew Thrift Sent: Donnerstag, 18. April 2013 06:44 To: Dietmar Maurer; pve-devel@pve.proxmox.com Subject: Re: [pve-devel] Add sound card to VM's from WebUI RDP still requires that the Terminal Server has a sound card. At least my win7 on a PVE servers play sounds (no physical sound card and no KVM sound card). I just connect with my win desktop rdp client. Can you explain this? Martin ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] Add sound card to VM's from WebUI
Is this feature planned ? And if not, would patches be accepted to add this feature ? Use cases: -Virtual Terminal Services (Required for RDS Remote Sound) -Virtual Desktop Infrastructure Based on the guest OS the default sound card could be selected, e.g. ac97 for 32bit WinXP, hda for 64bit Windows/Linux Regards, Andrew Thrift ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Process to submit patches
Hi Dietmar, We have been running this successfully for a few weeks now on the default PVE kernel with no issues. I searched the list archives, and the issues I could see looked like they were related to the MTU on the parent physical interface being too small. Linux does not clearly differentiate between L2 (interface) and L3 (IP) MTU, and also has a hidden 4 byte allowance on the interface MTU, making it confusing. When running QinQ the parent physical interface needs a MTU of at least 1504 bytes (allows a 1508 byte frame), any sub vlan's and bridges will then inherit this MTU allowing for the inner tag to pass externally correctly. Obviously if the users are not doing QinQ the parent physical interface can be the default 1500byte MTU and will work perfectly. We have only tested with KVM, so maybe there are other issues with OpenVZ... If there are not, it would be good to see this patch merged. Regards, Andrew On 4/2/2013 5:32 PM, Dietmar Maurer wrote: Our patches creates the vlan sub-if on the parent VM bridge, rather than on the parent interface. This works with both QinQ and non QinQ configurations. AFAIR this create serious problem. We already tried that several times and always run into problems. Search the list for details. ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] Process to submit patches
Hi, We are wanting to submit a patch to Network.pm to be included upstream. Our patch changes the way Proxmox dynamically creates vlans allowing for the current model, as well as for QinQ. Currently it is not possible to do QinQ on Proxmox as when you specify a VLAN in the WebUI, Proxmox checks the parent bridge for a Physical Interface, then creates the vlan sub-if on the Physical Interface and then binds it to a new bridge. Our patches creates the vlan sub-if on the parent VM bridge, rather than on the parent interface. This works with both QinQ and non QinQ configurations. What is the process to submit our patch ? Regards, Andrew Thrift ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Process to submit patches
FYI patch is: *** Network.pm.orig 2013-04-02 10:19:42.0 +1300 --- Network.pm 2013-04-02 11:41:44.0 +1300 *** *** 122,138 #check if we have an only one ethX or bondX interface in the bridge my $iface; ! PVE::Tools::dir_glob_foreach($dir, '((eth|bond)\d+)', sub { my ($slave) = @_; die more then one physical interfaces on bridge '$bridge'\n if $iface; ! $iface = $slave; }); die no physical interface on bridge '$bridge'\n if !$iface; ! my $ifacevlan = ${iface}.$tag; # create vlan on $iface is not already exist if (! -d /sys/class/net/$ifacevlan) { --- 122,138 #check if we have an only one ethX or bondX interface in the bridge my $iface; ! PVE::Tools::dir_glob_foreach($dir, '((eth|bond)\d+\.?\d+)', sub { my ($slave) = @_; die more then one physical interfaces on bridge '$bridge'\n if $iface; ! $iface = $bridge; }); die no physical interface on bridge '$bridge'\n if !$iface; ! my $ifacevlan = ${bridge}.$tag; # create vlan on $iface is not already exist if (! -d /sys/class/net/$ifacevlan) { On 4/2/2013 11:27 AM, Andrew Thrift wrote: Hi, We are wanting to submit a patch to Network.pm to be included upstream. Our patch changes the way Proxmox dynamically creates vlans allowing for the current model, as well as for QinQ. Currently it is not possible to do QinQ on Proxmox as when you specify a VLAN in the WebUI, Proxmox checks the parent bridge for a Physical Interface, then creates the vlan sub-if on the Physical Interface and then binds it to a new bridge. Our patches creates the vlan sub-if on the parent VM bridge, rather than on the parent interface. This works with both QinQ and non QinQ configurations. What is the process to submit our patch ? Regards, Andrew Thrift ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] QEMU/KVM + RBD latency
Hi, Any FYI more than anything. We have experienced VERY high latency when performing IO heavy operations on our RBD volumes. We initially thought it was due to network congestion on our uplinks, but it turns out it is a bug. See http://tracker.ceph.com/issues/3737 It would be great if Proxmox could include this when it is included in the stable branch of Ceph. Regards, Andrew ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] features ideas for 3.0 roadmap
Having a "Maintenance" mode would be good. Being able to mark a node as being under "Maintenance" and having all VM's migrate off it, and it disappear from the target list in migrations. Also being able to select Multiple guests and migrate them in one action would also be VERY useful. Regards, Andrew On 3/13/2013 8:21 AM, James Devine wrote: Perhaps some type of update mode which would pull a host server out of the cluster and migrate machines off of it. Auto-migrate VMs to do resource balancing. I've heard resistance to this with the argument that the migration would cause an even higher load but I think it is still a valid option. If the high load ends up being a long term situation it may be beneficial to migrate and cause a high load for a brief period of time. Also if things are kept balanced early enough the problem may be prevented all together. There could be tuning options such as when to balance and delay to prevent short periods of high load from causing a migration. This would also be useful for rebalancing the cluster if new hosts are added or if the above option is implemented and a host is re-added to the cluster. On Wed, Mar 6, 2013 at 3:18 AM, James A. Coyle james.co...@jamescoyle.net wrote: I would +1 the backups, it's the one part of Proxmox which drives me mad! I tend to SSH in and rename them manually for major backups. I'd like to see some small UI changes: - Add a search box under Storage Content so that we can start to type and the content is filtered. - On Create Backup Job - allow a name or comment to describe the backup. - As discussed - backups are tricky to manage on VMID alone. Overall though, it's a great product! I use 2 x 3 node clusters and use about a 50:50 split between QEMU and OpenVZ. Thanks for your efforts! James Coyle M:+44 (0) 7809 895 392 E: james.co...@jamescoyle.net Skype: jac2703 Gtalk: jac2...@gmail.com www: www.jamescoyle.net - Original Message - From: "FinnTux" finn...@gmail.com To: "Alexandre DERUMIER" aderum...@odiso.com Cc: pve-devel@pve.proxmox.com Sent: Wednesday, 6 March, 2013 7:52:56 AM Subject: Re: [pve-devel] features ideas for 3.0 roadmap - finish vm/templates copy/clone - storage migration - storage migration + live migration for local disks - add hyperv cpu features for windows 2008 - add spice console support (openstack has added spice-html5 console, so it should work now) - add support for qemu guest agent - finish cleanup of hotplug/unplug with cooperation of guest agent - add dynamic nic vlan change - add dynamic nic rate limit change What do you think about this ? Other ideas are welcome :) Sounds great. - Add vCPU, vRAM and bridges (networks) to pools - Improve backup handling. Like mentioned before on the forums backup name based solely on VMID is confusing. Needs somekind of notes included in the backup ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] network card MBit/s vs MB/s
I agree. All networking measurements are in bits/s Megabits, Kilobits Having Proxmox's in MB/s is inconsistent with the way the rest of the industry measures network bandwidth. Regards, Andrew On 2/17/2013 8:13 PM, Stefan Priebe - Profihost AG wrote: Right but people are used to use Mbit/s for everything network related. So I think a lot of people who want to use it will use it incorrect. Stefan Am 17.02.2013 um 07:43 schrieb Dietmar Maurer diet...@proxmox.com: nobody an opinion about that? To me it had happened several times that i putted in 100MB/s and i wanted 100Mbit/s. I guess not many people use that feature. But I prefer MB, because all other things (beside network) are measured in MB, for example HD or backup speed, disk size, ... ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] vm migration + storage migration with ndb workflow notes
As a user, our main use case for live storage migration would be migrating between local disk/RBD/iSCSI. We would also use the ability to migrate storage across the network. This could for example be used to migrate between a SAN in one datacentre to a SAN in a remote datacentre. Regards, Andrew Dietmar Maurer diet...@proxmox.com wrote: I have finally understand the whole workflow for migrate vm + storage migrate at the same time (I don't have to much time to work on this, maybe it's better to have stable working disk clone/copy/live mirror code before ) Yes, I would also like to do that after copy/clone. Besides, I think the whole thing gets to complex. We only want to be able to migrate local disks (changing storage is not really needed?). I guess that would make 'remote' migrate much easier? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel -- Sent from my Android phone with K-9 Mail. Please excuse my brevity.___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel