changelog:
- remove restart option
- check if vm|ct are running on a bridge delete
- run the networking service reload in a task
This add a new api to online reload networking configuration
with ifupdown2.
This work with native ifupdown2 modules, as ifupdown2 have
interface dependency
---
src/PVE/INotify.pm | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/PVE/INotify.pm b/src/PVE/INotify.pm
index 0b9ea4a..4024951 100644
--- a/src/PVE/INotify.pm
+++ b/src/PVE/INotify.pm
@@ -875,6 +875,8 @@ sub __read_etc_network_interfaces {
if (($id eq 'address')
Changelog:
- Various cleanups from Wolfgang comments
- add tests for vxlan, bridgeports && bridge options
This patches serie add support for new interfaces and options:
- mtu : for ifupdown && ifupdown2
- vxlan interfaces for ifupdown2
- bridgeports options for ifupdown2
Alexandre Derumier
---
src/PVE/INotify.pm | 47 ++-
test/etc_network_interfaces/t.create_network.pl | 51 +
2 files changed, 97 insertions(+), 1 deletion(-)
diff --git a/src/PVE/INotify.pm b/src/PVE/INotify.pm
index 4024951..97c4235 100644
>>is there any more documentation for those flags and why exactly they are
>>needed (the redhat bugzilla entry is rather sparse with information)
but some notes in kernel patch
https://patchwork.kernel.org/patch/7727711/
"Per Hyper-V specification (and as required by Hyper-V-aware guests),
>>I'm not sure that's a good idea. In theory `restart` would have worked
>>before as well, but the consequence in this case is *always* that the
>>API call fails due to a connection loss.
>>IMO we should only ever do 'reloads' - they too can kill the connection
>>of course, but restart
Hi,
I have found this presentation about coming corosync3 (seem to be alpha
recently)
http://build.clusterlabs.org/corosync/presentations/2017-Kronosnet-The-new-face-of-corosync-communications.pdf
with the new kronosnet (knet) transport.
Latencies results are really impressive and no more
* Combine sections from the wiki
* add section for avoiding RAID controllers
* correct command line for bluestore DB device creation
* minor rewording
Signed-off-by: Alwin Antreich
---
pveceph.adoc | 92 +++-
1 file changed, 60
On 6/26/18 3:38 PM, Stoiko Ivanov wrote:
> Signed-off-by: Stoiko Ivanov
> ---
> I like the codechanges in the 2 patches! - it's shorter and more clear than
> the
> previous version. - Thanks!
>
> One tiny thing which bit me, while trying it with the pveum changes was, that
> providing a prop as
Signed-off-by: Stoiko Ivanov
---
I like the codechanges in the 2 patches! - it's shorter and more clear than the
previous version. - Thanks!
One tiny thing which bit me, while trying it with the pveum changes was, that
providing a prop as $sort_key resulted in a warning - see this fixup.
One
this series adds some useful api calls that make use of the qemu-guest-agent
like
set-user-password
exec
file-read
file-write
it also adds some to qm
i plan to implement a 'qm push' and 'qm pull' in the future,
but i wanted to send the things i finished up until now,
so that they can get
this adds a command 'qm ga passwd' so that we can reuse
'qm ga' for future guest agent calls
Signed-off-by: Dominik Csapak
---
PVE/CLI/qm.pm | 6 ++
1 file changed, 6 insertions(+)
diff --git a/PVE/CLI/qm.pm b/PVE/CLI/qm.pm
index 42b8f2c..a7aaf93 100755
--- a/PVE/CLI/qm.pm
+++
this executes the guest agent command 'set-user-password'
with which one can change the password of an existing user in the vm
Signed-off-by: Dominik Csapak
---
PVE/API2/Qemu/Agent.pm | 62 +-
1 file changed, 61 insertions(+), 1 deletion(-)
diff
this api call reads a file via the guest agent,
(in 1MB chunks) but is limited to 16MiB (for now)
if the file is bigger, the output gets truncated and a
'truncated' flag is set in the return object
Signed-off-by: Dominik Csapak
---
PVE/API2/Qemu/Agent.pm | 76
on the commandline the implementation for exec is a bit different
because there we want (by default) to wait for the result,
as opposed to the api, where it is enough to return the pid and
let the client handle the polling
this behaviour is optional and can be turned off, as well as the
timeout
writes the given content to the file
the size is at the moment limited by the max post size of the
pveproxy/daemon, so we set the maxLength to 60k
Signed-off-by: Dominik Csapak
---
PVE/API2/Qemu/Agent.pm | 40
1 file changed, 40 insertions(+)
diff
this imitates the qemu-guest-agent interface
with an 'exec' api call which returns a pid
and an 'exec-status' api call which takes a pid
the command for the exec call is given as an 'alist'
which means that when using we have to give the 'command'
parameter multiple times e.g.
pvesh create
Signed-off-by: Dominik Csapak
---
PVE/API2/Qemu/Agent.pm | 5 ++--
PVE/QemuServer/Agent.pm | 65 +
PVE/QemuServer/Makefile | 1 +
3 files changed, 68 insertions(+), 3 deletions(-)
create mode 100644 PVE/QemuServer/Agent.pm
diff --git
On Fri, Jun 22, 2018 at 03:06:26AM +0200, Alexandre Derumier wrote:
> This add a new api to online reload networking configuration
> with ifupdown2.
>
> This work with native ifupdown2 modules, as ifupdown2 have
> interface dependency relationships.
>
> an optional "restart" param is available,
Now, we do not change the ordering by default.
Signed-off-by: Dietmar Maurer
---
src/PVE/CLIHandler.pm | 24
1 file changed, 16 insertions(+), 8 deletions(-)
diff --git a/src/PVE/CLIHandler.pm b/src/PVE/CLIHandler.pm
index dfdc41c..ec9c78d 100644
---
great :), i like this one much better
some problems though (sorry for not catching them earlier)
the checkbox should probably not be there when editing an existing storage
when editing a 'non-pve' ceph storage, it spits out the error:
Uncaught TypeError: this[c._config.names.set] is not a
This way we can access all type information and can use that to sort
numeric columns correctly.
Signed-off-by: Dietmar Maurer
---
src/PVE/CLIHandler.pm | 83 +++
1 file changed, 38 insertions(+), 45 deletions(-)
diff --git a/src/PVE/CLIHandler.pm
Combine both dialogues. This not only helps to reuse code but also
reduces storage choices from the Storage -> Add menu, and thus
improves usability.
Signed-off-by: Thomas Lamprecht
---
changes v1 -> v2:
* do not hide the monitor or username field when choosing the
hyperconverged variant but
* Adds list for all objects currently handled in the CLI
* Adds dump for User and Group
* Uses print_api_list and print_api_result from PVE::CLIHandler
Signed-off-by: Stoiko Ivanov
---
Changes from v5:
* adapt to the new methods in pve-common - supersedes only the last commit in
the series.
Signed-off-by: Stoiko Ivanov
---
Took a quick look at the patch - LGTM.
ran into a tiny problem, if we want to print a property, which is not defined
for an entry (e.g. comment for a user).
src/PVE/CLIHandler.pm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
On 06/24/2018 08:47 PM, Alexandre Derumier wrote:
Users have reported cpu problem with last windows 10 update (1803)
https://forum.proxmox.com/threads/high-cpu-load-for-windows-10-guests-when-idle.44531/
adding hyperv hv_stimer && hv_synic fix the problem.
This patch series add support to
On 6/25/18 1:31 PM, Dominik Csapak wrote:
> the code looks good and i'm all for doing this,
thanks for the look!
> but i do not really like how the endresult looks
>
> an improvement would be to put the checkbox more at the top, or in the footer
> area?
> it feels rather weird to have a
27 matches
Mail list logo