With recent changes, pvebackup_co_run_next_job cancels the job async,
so we need to run pvebackup_co_cleanup in the completion handler
instead. We call pvebackup_co_run_next as long as there are
jobs in the list.
Signed-off-by: Dietmar Maurer
---
blockdev.c | 9 -
1 file changed, 4 inser
The destroy_vm in the eval could only die if the VM were a template.
This is not possible here.
Signed-off-by: Dominic Jäger
---
PVE/CLI/qm.pm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/PVE/CLI/qm.pm b/PVE/CLI/qm.pm
index a378d3d..a16033c 100755
--- a/PVE/CLI/qm.pm
+++ b
As mentioned on the mailing list [0] disks owned by the VM and unused
disks should be removed before the config file is removed.
[0] https://pve.proxmox.com/pipermail/pve-devel/2019-October/039593.html
Signed-off-by: Dominic Jäger
---
PVE/QemuServer.pm | 12 ++--
1 file changed, 6 inser
The variable was used in one place only. Nesting the code into the
lock_config_full makes clear what code is locked.
Signed-off-by: Dominic Jäger
---
PVE/CLI/qm.pm | 11 +++
1 file changed, 3 insertions(+), 8 deletions(-)
diff --git a/PVE/CLI/qm.pm b/PVE/CLI/qm.pm
index c93f78d..a378d3d
The whole $importfn is executed in a lock_config_full. Consequently,
1. lock_config is redundant
2. it is not possible that a lock has been written in the VM's config file
(check_lock) in the meanwhile
3. it is not possible that the VM has been started (check_running) in the
meanwhile
Signe
Functions like importovf can now use a parameter lock in the config.
Signed-off-by: Dominic Jäger
---
PVE/CLI/qm.pm| 4 ++--
PVE/QemuServer/ImportDisk.pm | 6 --
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/PVE/CLI/qm.pm b/PVE/CLI/qm.pm
index a16033c..97f0f8
Previously creating a VM on another node between locking the config with
lock_config_full and writing to it for the first time with write_config
was possible.
Using create_and_lock_config eliminates this possibility. This means
that now a parameter lock in the config is used instead of flock only.
This series cleans up some redundant locks and functions and sets more
appropriate locks instead when importing .ovf and disks.
Dominic Jäger (7):
Move config removal after disk removal
Remove vm_destroy
Remove redundant locks
Remove variable from lock
Remove useless eval
Add skiplock
This function has been used in one place only. Removing it avoids confusion
between vm_destroy and vm_destroy.
Signed-off-by: Dominic Jäger
---
I copied all the checks from vm_destroy even if they were useless.
Removing those useless functions is a follow up patch.$
PVE/CLI/qm.pm | 11 +
On 10/25/19 11:24 AM, Dominic Jäger wrote:
> As mentioned on the mailing list [0] disks owned by the VM and unused
> disks should be removed before the config file is removed.
>
> [0] https://pve.proxmox.com/pipermail/pve-devel/2019-October/039593.html
>
> Signed-off-by: Dominic Jäger
> ---
> P
On Fri, Oct 25, 2019 at 11:49:51AM +0200, Thomas Lamprecht wrote:
> On 10/25/19 11:24 AM, Dominic Jäger wrote:
> > As mentioned on the mailing list [0] disks owned by the VM and unused
> > disks should be removed before the config file is removed.
> >
> > [0] https://pve.proxmox.com/pipermail/pve-
to improve UX, disabled child fields shouldn't show validation errors
when the parent widget (radiofield) isn't selected anymore.
Signed-off-by: Tim Marx
---
changes v2:
* swapped ternary for if/else
* make use of new cdImageField var
* adapted commit message
www/manager6/qemu/CDEdit.js | 9 +++
Prevent the comboBox from displaying a validation error although
forceSelection is true. If you change a valid selection by removing
characters manually and click somewhere else, the comboBox restores the
selection with the previous value. The validation logic then checked the
restored value, but c
with qemu 4.0.1, there is now a machine type pc-q35-4.0.1 which does not fit
into our regex
this broke live migration of q35, as we give the machine type (incl version
info) to 'qm start' on the target node, which checks it against the
JSONSchema
to fix this, extend the regex to allow any number
On 10/25/19 2:36 PM, Dominik Csapak wrote:
> with qemu 4.0.1, there is now a machine type pc-q35-4.0.1 which does not fit
> into our regex
>
> this broke live migration of q35, as we give the machine type (incl version
> info) to 'qm start' on the target node, which checks it against the
> JSONSch
The libanyevent-perl version 7.140-3 included a fix for this.
It migrated to the then still testing (buster was not yet released)
on 07.04.2019, and so we can safely revert this workaround again
here.
Albeit this was fixed since Buster was officially released, still
bump the version dependency to
On 10/22/19 4:40 PM, Christian Ebner wrote:
> Signed-off-by: Christian Ebner
> ---
> version 5:
>* only show checkbox for CT/VM destroy dialog (as suggested)
>* added qtip to checkbox
>
> www/manager6/window/SafeDestroy.js | 22 ++
> 1 file changed, 22 insertions(+)
On 10/25/19 12:47 PM, Tim Marx wrote:
> to improve UX, disabled child fields shouldn't show validation errors
> when the parent widget (radiofield) isn't selected anymore.
>
> Signed-off-by: Tim Marx
> ---
> changes v2:
> * swapped ternary for if/else
> * make use of new cdImageField var
> * adap
On 10/25/19 1:06 PM, Tim Marx wrote:
> Prevent the comboBox from displaying a validation error although
> forceSelection is true. If you change a valid selection by removing
> characters manually and click somewhere else, the comboBox restores the
> selection with the previous value. The validation
On 10/25/19 11:24 AM, Dominic Jäger wrote:
> This function has been used in one place only. Removing it avoids confusion
> between vm_destroy and vm_destroy.
>
> Signed-off-by: Dominic Jäger
> ---
> I copied all the checks from vm_destroy even if they were useless.
> Removing those useless functi
On 10/25/19 11:24 AM, Dominic Jäger wrote:
> The destroy_vm in the eval could only die if the VM were a template.
> This is not possible here.
>
> Signed-off-by: Dominic Jäger
> ---
> PVE/CLI/qm.pm | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/PVE/CLI/qm.pm b/PVE/CLI/
"remove variable from lock" is as generic and obscure as it gets,
could be anything ^^ Please use a real commit subject.. concise and
intended for human digestions (e.g., no full paths, if not an essential
part for human understanding)
On 10/25/19 11:24 AM, Dominic Jäger wrote:
> The variable was
On 10/25/19 11:24 AM, Dominic Jäger wrote:
> Previously creating a VM on another node between locking the config with
> lock_config_full and writing to it for the first time with write_config
> was possible.
>
> Using create_and_lock_config eliminates this possibility. This means
> that now a par
23 matches
Mail list logo
