On 11/20/19 11:35 AM, Tim Marx wrote:
The bugfix for #2317 introduced a kind of odd api behavior, where each volume
was returned twice from our api if a storage has both 'rootdir' & 'images'
content
types enabled. To give the content type of the volume an actual meaning, it is
now inferred form
On 11/20/19 3:49 PM, Oguz Bektas wrote:
> this allows pending changes to be applied when a reboot is issued inside
> the container.
>
> Signed-off-by: Oguz Bektas
> ---
> src/lxc-pve-prestart-hook | 9 -
> 1 file changed, 8 insertions(+), 1 deletion(-)
>
> diff --git
On 11/19/19 3:25 PM, Matt Dunwoodie wrote:
> This adds an extra field to the AgentFeatureSelector that reflects the
> change in qemu-server.
>
> Changes since previous version:
>
note: those are really good, thanks for that, but normally the git history
doesn't has to care about review revision
On 11/19/19 12:23 PM, Stefan Reiter wrote:
> This series refactors QemuServer and creates three new packages:
> * 'PVE::QemuServer::Helpers' for general purpose helpers
> * 'PVE::QemuServer::Monitor' for higher-level QMP functions
> * 'PVE::QemuServer::Machine' for QEMU machine-type related
On 11/19/19 1:25 PM, Thomas Lamprecht wrote:
> Add a simple ExecStartPre command which reads the local node config,
> and if a delay is set the helper sleeps that long then exists.
>
> The systemd-unit approach was chosen as this ensures that we really
> only delay when doing the startall on node
On 11/20/19 3:49 PM, Oguz Bektas wrote:
> pve-container:
>
> v3->v4:
> * merged commits
>
> note: left the prestart hook commit separate on purpose, since i think it's a
> different change (doesn't use the new reboot API or anything related with it)
>
yeah, seems valid to me. As it affects
On 10/31/19 1:36 PM, Dominik Csapak wrote:
> this series lets users add tags to guest configs
> these do not have any concrete meaning but are intended to be used
> by management software such as config management systems
>
> this is the basic implementation of this feature, in the next step
> i
On 11/20/19 8:30 AM, Wolfgang Bumiller wrote:
> Changes to v3:
> * Change mount point staging directory from /run/pve/mountpoints to
> /var/lib/lxc/.pve-staged-mounts due to allowed mount paths being
> restricted by the lxc-start apparmor profile.
> (Only affects path 8, the remaining
On 11/12/19 2:23 PM, Dominik Csapak wrote:
> new pve-common breaks old manager and qemu-server
but that isn't true or, as we don't break for GUI changes
(cannot do a intra-cluster break) and common was kept backward
compatible, or?
I bumped it without breaks, will do the dependency versioned
On 11/12/19 2:23 PM, Dominik Csapak wrote:
> but prepend '' to ids where no domain is given, to keep the ability
> to use the shorthand syntax (e.g. 00:01.0 instead of :00:01.0)
>
> Signed-off-by: Dominik Csapak
> ---
> src/PVE/SysFSTools.pm | 20 +---
> 1 file changed,
This series allows the API to handle all 8 possible links that corosync 3
supports. Additional verifications are added to support a smarter fallback
behaviour and abort join if the current cluster config is invalid or
unsupported.
verify_conf is not just nice-to-have, but required for patch 3
It does some basic sanity checking, warns the user about encryption
settings and unresolved hostnames, and finally makes sure that all nodes
have the same links configured (as well as comparing the configured
links to specified interfaces, if there are any).
A corosync.conf that has been created
add_corosync_link_properties/extract_corosync_link_args are introduced
as helpers to avoid hardcoding links in parameters=>properties on
several occasions, while still providing autocompletion with pvecm by
being seperate parameters instead of an array.
Maximum number of links is given as
Verify that the config of the new node is valid and compatible with the
cluster (i.e. that the links for the new node match the currently
configured nodes).
Additionally, fallback is provided via a new parameter to addnode,
'new_node_ip'. Previously, fallback was handled on the joining node, by
Adds API call GET /cluster/config/apiversion to retrieve remote clusters
join-API version (0 is assumed for versions without this endpoint).
Warn user if remote version differs, and select new fallback method only
if available. This ensures full compatibility between nodes/clusters
with and
Same as we do in assert_joinable, a cluster with unconfigured IPs will
fail start after creation anyway.
Make "check_ip" a standalone sub ("check_ip_configured") and improve
error messages all around.
Also move call to create_conf up, so if it fails we exit before SSH and
auth-key setup.
On 11/20/19 1:24 PM, Stefan Reiter wrote:
> Signed-off-by: Stefan Reiter
> ---
> PVE/QemuServer.pm | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
applied, thanks!
___
pve-devel mailing list
pve-devel@pve.proxmox.com
On 11/20/19 1:24 PM, Stefan Reiter wrote:
> to allow 'qm showcmd --snapshot ...' to show the correct machine type.
>
> Signed-off-by: Stefan Reiter
> ---
> PVE/QemuServer.pm | 6 +-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
>
Signed-off-by: Stefan Reiter
---
I tested 4.1.0 seperately to 4.1.1 for a while, so I left the patches in the
order I did them (i.e. 4.1 rebase first, then rebase that to 4.1.1). Patches 1-3
could technically be squashed together into a single 4.1.1 rebase.
Tested with backups/restores,
pve-container:
v3->v4:
* merged commits
note: left the prestart hook commit separate on purpose, since i think it's a
different change (doesn't use the new reboot API or anything related with it)
Oguz Bektas (2):
implement container reboot
apply pending changes in lxc prestart hook
this allows pending changes to be applied when a reboot is issued inside
the container.
Signed-off-by: Oguz Bektas
---
src/lxc-pve-prestart-hook | 9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/src/lxc-pve-prestart-hook b/src/lxc-pve-prestart-hook
index
move stopBtn into shutdownBtn as a menu item.
we can remove the setDisabled() call for stopBtn near the end, since when
shutdownBtn is disabled, so is stopBtn.
Signed-off-by: Oguz Bektas
---
www/manager6/lxc/Config.js | 25 +++--
1 file changed, 11 insertions(+), 14
to right-click menu and dropdown menu
Signed-off-by: Oguz Bektas
---
www/manager6/lxc/CmdMenu.js | 16
www/manager6/lxc/Config.js | 9 +
2 files changed, 25 insertions(+)
diff --git a/www/manager6/lxc/CmdMenu.js b/www/manager6/lxc/CmdMenu.js
index 579b2585..e922546a
* add helper function vm_reboot
* add vm_reboot api call
* add 'pct reboot'
Signed-off-by: Oguz Bektas
---
src/PVE/API2/LXC/Status.pm | 52 ++
src/PVE/CLI/pct.pm | 3 ++-
src/PVE/LXC.pm | 13 ++
3 files changed, 67 insertions(+),
* Fix VMA tool build
* Change PVE code to new blockjob API
* Acquire missing lock for block_job_add_bdrv
Signed-off-by: Stefan Reiter
---
First two patches by Thomas.
Last one added by me, otherwise Backup immediately crashes with 'Permission
denied' in 4.1. My patch makes it work, but I'm
Signed-off-by: Stefan Reiter
---
Based on qemu refactor series[0], especially patch 6, can send a version for
current master as well, but then the refactor has to be rebased because of
qemu_machine_version_enabled -> Helpers::min_version.
Tested with snapshot create/restore between 3.0,
Signed-off-by: Stefan Reiter
---
...-PVE-Deprecated-adding-old-vma-files.patch | 46 +++
qemu | 2 +-
2 files changed, 29 insertions(+), 19 deletions(-)
diff --git a/debian/patches/pve/0021-PVE-Deprecated-adding-old-vma-files.patch
On 11/19/19 4:18 PM, Aaron Lauterer wrote:
> with pve-qemu-4.0.1-3 or higher it was not possible in a spice remote
> session to enable more displays on the fly in linux guests.
>
> Adding the `max_outputs` parameter to the qxl device manually restores
> the functionality.
>
> Signed-off-by:
On 11/20/19 3:38 PM, Thomas Lamprecht wrote:
> found no use with tree-wide search, so remove:
> * nic_models
> * os_list_description
>
> Both were introduced before the import to SVN happened.
>
> Signed-off-by: Thomas Lamprecht
> ---
>
> holler me if you find a use
.. as this is already
found no use with tree-wide search, so remove:
* nic_models
* os_list_description
Both were introduced before the import to SVN happened.
Signed-off-by: Thomas Lamprecht
---
holler me if you find a use
PVE/QemuServer.pm | 22 --
1 file changed, 22 deletions(-)
diff --git
On Wed, Nov 20, 2019 at 02:33:29PM +0100, Thomas Lamprecht wrote:
> On 11/19/19 6:37 PM, Oguz Bektas wrote:
> > also use the opportunity to refactor the shutdown button code into the
> > menu.
> >
> > Signed-off-by: Oguz Bektas
> > ---
> >
> > v2->v3:
> > no changes.
> >
> >
On 11/19/19 6:36 PM, Oguz Bektas wrote:
> i'm omitting the widget-toolkit patch since that one is already applied.
>
>
> pve-container:
> v2->v3:
> * use vmconfig_apply_pending() in prestart hook instead of calling
> update_pct_config and update_lxc_config
>
> Oguz Bektas (4):
> add vm_reboot
looks like all the issues we talked about are fixed now :)
nice work
Tested-by: Oguz Bektas
On Wed, Nov 20, 2019 at 08:30:55AM +0100, Wolfgang Bumiller wrote:
> Changes to v3:
> * Change mount point staging directory from /run/pve/mountpoints to
> /var/lib/lxc/.pve-staged-mounts due to
to allow 'qm showcmd --snapshot ...' to show the correct machine type.
Signed-off-by: Stefan Reiter
---
PVE/QemuServer.pm | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index 11e7169..e6369c1 100644
--- a/PVE/QemuServer.pm
+++
Signed-off-by: Stefan Reiter
---
PVE/QemuServer.pm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index e6369c1..4f968c3 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -3905,7 +3905,7 @@ sub config_to_command {
my
The bugfix for #2317 introduced a kind of odd api behavior, where each volume
was returned twice from our api if a storage has both 'rootdir' & 'images'
content
types enabled. To give the content type of the volume an actual meaning, it is
now inferred form the associated guest (qemu/lxc) and we
Signed-off-by: Tim Marx
---
PVE/Storage/Plugin.pm | 20 ++--
1 file changed, 10 insertions(+), 10 deletions(-)
diff --git a/PVE/Storage/Plugin.pm b/PVE/Storage/Plugin.pm
index 2bf6f0a..c2612c1 100644
--- a/PVE/Storage/Plugin.pm
+++ b/PVE/Storage/Plugin.pm
@@ -938,21 +938,21 @@
Signed-off-by: Tim Marx
---
PVE/CLI/pvesm.pm | 12 +++-
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/PVE/CLI/pvesm.pm b/PVE/CLI/pvesm.pm
index 01ddd23..63b212a 100755
--- a/PVE/CLI/pvesm.pm
+++ b/PVE/CLI/pvesm.pm
@@ -124,21 +124,23 @@ my $print_content = sub {
my
On Wed, Nov 20, 2019 at 09:09:12AM +0100, Fabian Ebner wrote:
> On 11/19/19 12:05 PM, Tim Marx wrote:
> > The bugfix for #2317 introduced a kind of odd api behavior, where each
> > volume
While Tim's patch does indeed fix the API, I agree with Fabian's objections:
> Isn't it necessary to do
On 11/20/19 8:31 AM, Wolfgang Bumiller wrote:
Signed-off-by: Wolfgang Bumiller
---
Changes to v3:
Changed the staging path from /run/pve/mountpoints to
/var/lib/lxc/.pve-staged-mounts due to the lxc-start apparmor profile.
src/PVE/LXC.pm | 30 +-
1 file
On 11/19/19 12:05 PM, Tim Marx wrote:
The bugfix for #2317 introduced a kind of odd api behavior, where each volume
was returned twice from our api if a storage has both 'rootdir' & 'images'
content
types enabled. To give the content type of the volume an actual meaning, it is
now inferred form
41 matches
Mail list logo