On Sat, Apr 04, 2020 at 08:09:58PM +0200, Thomas Lamprecht wrote:
> On 4/3/20 4:37 PM, w.bumil...@proxmox.com wrote:
> > ...
> > +# Change the cpu "shares" for a container.
> > +#
> > +# In cgroupv1 we used a value in `[0..50]` with a default of 1024.
> > +#
> > +# In cgroupv2 we do not have "s
As it might die with an error which should end up in the
_exit() code path rather than bailing out into the upper
scope.
Signed-off-by: Wolfgang Bumiller
---
src/PVE/RESTEnvironment.pm | 18 ++
1 file changed, 10 insertions(+), 8 deletions(-)
diff --git a/src/PVE
$resfh can be a pipe from POSIX::pipe() or the upid output
handle, which is an IO::File, so we need to take its
fileno().
Signed-off-by: Wolfgang Bumiller
Fixes: ed52a8435a6d ("fork_worker: use separate pipe for status messages")
---
Fixes 'Error: unable to read tail'
The backend's minimum value for container memory is 16M,
not 32.
Signed-off-by: Wolfgang Bumiller
---
www/manager6/lxc/ResourceEdit.js | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/www/manager6/lxc/ResourceEdit.js b/www/manager6/lxc/ResourceEdit.js
index 6bd35e8a..641
On Mon, Dec 25, 2017 at 09:44:52AM +0100, c.mo...@web.de wrote:
> The avg. memory usage of LXC running HAProxy is 10MB according to WebUI.
> Checking the amount of free and used system memory in LXC shows:
> vm102-haproxy:~# free
> total used free shared buffers cach
applied
On Thu, Dec 21, 2017 at 12:52:53PM +0100, Dominik Csapak wrote:
> we do not use the applet anymore, and setting it throws an error
> in the gui when clicking the console button
>
> map it to 'html5' and mark it deprecated, so that we can remove
> that setting in the next major release
>
applied
On Fri, Dec 22, 2017 at 02:34:30PM +0100, Thomas Lamprecht wrote:
> remove useless sleep on stop and make pmxcfs restart fully transparent
> for IPCC users, at least if successful.
>
> Not only retry once if connection cache gets trashed but also poll for
> a connection if a pmxcfs restar
On Wed, Dec 20, 2017 at 03:13:36PM +0100, Thomas Lamprecht wrote:
> Signed-off-by: Thomas Lamprecht
> ---
> src/PVE/JSONSchema.pm | 6 ++
> 1 file changed, 6 insertions(+)
>
> diff --git a/src/PVE/JSONSchema.pm b/src/PVE/JSONSchema.pm
> index 9c26184..2b646fd 100644
> --- a/src/PVE/JSONSchem
Signed-off-by: Wolfgang Bumiller
Fixes: 2bfbee039bbe ("include format for efidisk")
---
PVE/QemuServer.pm | 10 --
1 file changed, 8 insertions(+), 2 deletions(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index 42c412a..4837331 100644
--- a/PVE/QemuServer.pm
applied whole series
On Wed, Dec 20, 2017 at 11:34:01AM +0100, Fabian Grünbichler wrote:
> otherwise it gets confused and enables pve-ha-crm twice in the postinst.
>
> Signed-off-by: Fabian Grünbichler
> ---
> Note: the behaviour change is not related to the removed blank line
>
> debian/rules
On Mon, Dec 18, 2017 at 10:21:38AM +0100, Thomas Lamprecht wrote:
> move variables nearer to where they actually used.
> drop program name early from argv array
> drop unnecessary variables
>
> Signed-off-by: Thomas Lamprecht
> ---
> src/PVE/CLIHandler.pm | 39 +++
On Mon, Dec 18, 2017 at 10:21:40AM +0100, Thomas Lamprecht wrote:
> reduce code reuse and prepare for sub commands
>
> Signed-off-by: Thomas Lamprecht
> ---
> src/PVE/CLIHandler.pm | 138
> --
> 1 file changed, 77 insertions(+), 61 deletions(-)
>
On Mon, Dec 18, 2017 at 10:21:41AM +0100, Thomas Lamprecht wrote:
> allow to use sub commands alá
> # pveum user add
>
> The new resolve_cmd traverses $cmddef, resolves one level of aliases
> and returns the respective sub command, its cmddef, arguments and if
> it was expanded (e.g., pveum u d .
On Mon, Dec 18, 2017 at 10:21:40AM +0100, Thomas Lamprecht wrote:
> reduce code reuse and prepare for sub commands
>
> Signed-off-by: Thomas Lamprecht
> ---
> src/PVE/CLIHandler.pm | 138
> --
> 1 file changed, 77 insertions(+), 61 deletions(-)
>
applied
On Mon, Dec 18, 2017 at 10:21:35AM +0100, Thomas Lamprecht wrote:
> passing one param less to the helper methods
>
> Signed-off-by: Thomas Lamprecht
> Reviewed-by: Dominik Csapak
> ---
> src/PVE/CLIHandler.pm | 29 -
> 1 file changed, 12 insertions(+), 17 de
applied
On Mon, Dec 18, 2017 at 10:21:36AM +0100, Thomas Lamprecht wrote:
> Signed-off-by: Thomas Lamprecht
> Reviewed-by: Dominik Csapak
> ---
> src/PVE/CLIHandler.pm | 16 +++-
> 1 file changed, 11 insertions(+), 5 deletions(-)
>
> diff --git a/src/PVE/CLIHandler.pm b/src/PVE/CLI
applied
On Mon, Dec 18, 2017 at 10:21:37AM +0100, Thomas Lamprecht wrote:
> use shorter and also a bit faster methods to expand and get comand
> names
>
> Signed-off-by: Thomas Lamprecht
> ---
>
> src/PVE/CLIHandler.pm | 25 +++--
> 1 file changed, 3 insertions(+), 22 delet
applied
On Mon, Dec 18, 2017 at 10:21:39AM +0100, Thomas Lamprecht wrote:
> will be reused in later patches too
>
> Signed-off-by: Thomas Lamprecht
> Reviewed-by: Dominik Csapak
> ---
> src/PVE/CLIHandler.pm | 23 ---
> 1 file changed, 12 insertions(+), 11 deletions(-)
>
>
delete"
>
> Signed-off-by: Thomas Lamprecht
once the rest is finished:
Acked-by: Wolfgang Bumiller
> ---
> src/PVE/CLIHandler.pm | 14 ++
> 1 file changed, 6 insertions(+), 8 deletions(-)
>
> diff --git a/src/PVE/CLIHandler.pm b/src/PVE/CLIHandle
On Mon, Dec 18, 2017 at 10:21:43AM +0100, Thomas Lamprecht wrote:
> Signed-off-by: Thomas Lamprecht
> Reviewed-by: Dominik Csapak
once the rest is finished:
Acked-by: Wolfgang Bumiller
> ---
> src/PVE/CLIHandler.pm | 23 +++
> 1 file changed, 23 insertio
Signed-off-by: Wolfgang Bumiller
Fixes: 2bfbee039bbe ("include format for efidisk")
---
Changes since v1:
* When the disk was specified as a path instead of as storage:volume
the format property is now mandatory.
PVE/QemuServer.pm | 11 +--
1 file changed, 9 insertions(+), 2
Currently this only allows specifying '+pcid' or '-pcid'
but might be extended in the future.
Signed-off-by: Wolfgang Bumiller
---
PVE/QemuServer.pm | 14 ++
1 file changed, 14 insertions(+)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index 32bf70d..3e3c
applied
On Mon, Jan 08, 2018 at 03:25:07PM +0100, Wolfgang Bumiller wrote:
> Signed-off-by: Wolfgang Bumiller
> Fixes: 2bfbee039bbe ("include format for efidisk")
> ---
> Changes since v1:
> * When the disk was specified as a path instead of as storage:volume
>
applied
On Tue, Jan 09, 2018 at 04:27:39PM +0100, Dominik Csapak wrote:
> but not in wizard, and default is off
> (because we do not know if the cpu supports it)
>
> Signed-off-by: Dominik Csapak
> ---
> www/manager6/qemu/CreateWizard.js | 1 +
> www/manager6/qemu/ProcessorEdit.js | 20 ++
applied both
On Wed, Dec 27, 2017 at 11:15:57AM +0100, Wolfgang Bumiller wrote:
> $resfh can be a pipe from POSIX::pipe() or the upid output
> handle, which is an IO::File, so we need to take its
> fileno().
>
> Signed-off-by: Wolfgang Bumiller
> Fixes: ed52a8435a6d ("fo
On Tue, Jan 09, 2018 at 06:42:45PM +0100, Alexandre Derumier wrote:
> Hugepages can take some time to be allocated by qemu at start (60s for 120G
> of 1G hugepages).
> This patch increase start timeout to 5min when hugepages are enabled.
Applied, but I don't feel too awesome about it.
OTOH a heur
applied and added another commit to indent new lines by 2 spaces:
- $cmdline =~ s/ -/ \\\n-/g if $param->{pretty};
+ $cmdline =~ s/ -/ \\\n -/g if $param->{pretty};
Now I can drop my sed wrapper ;-)
On Tue, Jan 09, 2018 at 09:56:28AM +0100, Thomas Lamprecht wrote:
> Shows each parame
applied
On Thu, Jan 11, 2018 at 09:49:31AM +0100, Dominik Csapak wrote:
> this was missing and lead to 'use of unitialized value $realm...'
> and a not working second server if a bindpw was defined
>
> Signed-off-by: Dominik Csapak
> ---
> PVE/Auth/LDAP.pm | 2 +-
> 1 file changed, 1 insertion(
applied
On Wed, Dec 13, 2017 at 11:34:48AM +0100, Dominik Csapak wrote:
> we had it previously on 8GB for containers, but with the refactoring,
> this got lost. this patch changes this back
>
> Signed-off-by: Dominik Csapak
> ---
> www/manager6/form/DiskStorageSelector.js | 6 +-
> www/mana
applied
___
pve-devel mailing list
pve-devel@pve.proxmox.com
https://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
applied
On Mon, Jan 08, 2018 at 09:38:52AM +0100, Thomas Lamprecht wrote:
> Signed-off-by: Thomas Lamprecht
> ---
>
> changes v1 -> v2:
> * remove anchors, their already applied on parsing
> * do not allow double colon separation,
> Net::SSLeay::X509_get_fingerprint only emits single colon sep
On Fri, Jan 05, 2018 at 10:37:38AM +0100, Dominik Csapak wrote:
> any comment ?
This is fine for running containers, but for stopped containers we
should rather remove the percentage & bar again instead of saying it's
using 0% of the disk. For CPU/Memory/Swap it makes sense that a stopped
containe
On Mon, Jan 15, 2018 at 08:30:33AM +0100, Alexandre DERUMIER wrote:
> >>But I thought you want to implement those checks directly in the GUI?
>
> Well, yes, but I would like to launch the migrateall, after before each
> migrate task, do the check if target host have enough memory.
>
> I don't th
applied both patches to master & stable-4
On Tue, Jan 16, 2018 at 10:21:41AM +0100, Fabian Grünbichler wrote:
> this have the 'spec-ctrl' flag by default to allow IBRS based Spectre
> mitigation by the guest kernel.
>
> Signed-off-by: Fabian Grünbichler
> ---
> Note: these require pve-qemu-kvm >
applied to master & stable-4
On Tue, Jan 16, 2018 at 10:06:41AM +0100, Fabian Grünbichler wrote:
> to allow VMs to use IBRS/IBPB on systems with patched microcode
>
> Signed-off-by: Fabian Grünbichler
> ---
> for master & stable-4, but the latter needs a pve-qemu-kvm backport first
>
> PVE/Qem
applied
On Tue, Jan 16, 2018 at 10:06:42AM +0100, Fabian Grünbichler wrote:
> Signed-off-by: Fabian Grünbichler
> ---
> qm.adoc | 60 +++-
> 1 file changed, 43 insertions(+), 17 deletions(-)
>
> diff --git a/qm.adoc b/qm.adoc
> index 75617
applied whole series
On Tue, Jan 09, 2018 at 01:25:28PM +0100, Thomas Lamprecht wrote:
> v3 of the sub command for CLI series. v2 can be found here:
> https://pve.proxmox.com/pipermail/pve-devel/2017-December/029916.html
>
> This are the remaining, not yet applied, patches from the sub command
>
applied
On Wed, Jan 17, 2018 at 10:33:19AM +0100, Dominik Csapak wrote:
> use the pveupgrade command directly without bash inbetween,
> the incorrect quoting led to '--shell' not being passed to
> pveupgrade and closing the connection
>
> Signed-off-by: Dominik Csapak
> ---
> changes from v2:
>
applied
On Mon, Jan 15, 2018 at 11:15:54AM +0100, Thomas Lamprecht wrote:
> Fixes a problem where a non root@pam system administrator does not
> sees the 'Permissions' tab for a storage in our WebUI.
>
> Fixes commit a2c18811d33d7e09765a7b0f09bba47bc9523822
>
> Signed-off-by: Thomas Lamprecht
>
applied
On Mon, Jan 22, 2018 at 11:00:07AM +0100, Thomas Lamprecht wrote:
> For sub commands we resolve the real $cmd, $def and its arguments,
> thus we should also get the handler from the resolved $def, not the
> global one.
>
> No change for normal (consisting of only the first argument)
> com
applied both patches
On Mon, Jan 22, 2018 at 12:12:40PM +0100, Dietmar Maurer wrote:
> Signed-off-by: Dietmar Maurer
> ---
> src/PVE/INotify.pm | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/src/PVE/INotify.pm b/src/PVE/INotify.pm
> index a383040..1e5687b 100644
>
applied all patches
On Mon, Jan 22, 2018 at 10:52:10AM +0100, Thomas Lamprecht wrote:
> Fixes problems in CLIHandler using the code pattern:
>
> while (my $line = <>) {
> ...
> }
>
> For why this causes only _now_ problems lets first look how <>
> behaves:
>
> "The null filehandle <> is spe
applied both patches
On Tue, Jan 23, 2018 at 08:19:55AM +0100, Thomas Lamprecht wrote:
> ---
>
> changes v1 -> v2:
> * do not export SOURCE_DATE_EPOCH, already done by
> /usr/share/dpkg/pkg-info.mk
>
> Makefile | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/Makefi
applied
On Thu, Jan 18, 2018 at 09:31:27AM +0100, Thomas Lamprecht wrote:
> Signed-off-by: Thomas Lamprecht
> ---
>
> changes v1 -> v2:
> * incorporate Fabian's comments (thanks!)
>
> pvecm.adoc | 14 ++
> 1 file changed, 14 insertions(+)
>
> diff --git a/pvecm.adoc b/pvecm.adoc
>
applied
On Wed, Jan 17, 2018 at 09:25:40AM +0100, Thomas Lamprecht wrote:
> For VMs we wait 180 seconds (CT sxtill 60, so no need to correct
> there) for stop to succeed. Reword and add a few details while
> touching this.
>
> Signed-off-by: Thomas Lamprecht
> ---
> qm.adoc | 14 --
applied with minor fixup
On Tue, Oct 17, 2017 at 02:11:49PM +0200, Alwin Antreich wrote:
> Signed-off-by: Alwin Antreich
> ---
> pvecm.adoc | 4
> 1 file changed, 4 insertions(+)
>
> diff --git a/pvecm.adoc b/pvecm.adoc
> index 36d18c2..11c6849 100644
> --- a/pvecm.adoc
> +++ b/pvecm.adoc
applied
On Tue, Jan 16, 2018 at 03:43:18PM +0100, Dominik Csapak wrote:
> we already encode the description in 'write_config',
> we do not need to do it here again
>
> this lead to newlines also to be encoded instead of
> splitting the description into lines and decode that
>
> Signed-off-by: Do
We don't use them and newer iproute2 doesn't like ':1' as
abbreviation for '0:1' apparently.
Signed-off-by: Wolfgang Bumiller
---
src/PVE/Network.pm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/PVE/Network.pm b/src/PVE/Network.pm
index
Forgot to include:
Fix #1637
On Mon, Jan 29, 2018 at 10:52:39AM +0100, Wolfgang Bumiller wrote:
> We don't use them and newer iproute2 doesn't like ':1' as
> abbreviation for '0:1' apparently.
>
> Signed-off-by: Wolfgang Bumiller
> ---
>
There's no technical reason for *us* to limit this to 120.
Signed-off-by: Wolfgang Bumiller
---
src/PVE/JSONSchema.pm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/PVE/JSONSchema.pm b/src/PVE/JSONSchema.pm
index c09a9f4..4bc7cbb 100644
--- a/src/PVE/JSONSchema.pm
applied, despite the lack of tab completion ;-)
On Fri, Jan 26, 2018 at 03:25:26PM +0100, Dominik Csapak wrote:
> and make them available with the up/down arrow key
>
> Signed-off-by: Dominik Csapak
> ---
> www/manager6/qemu/Monitor.js | 45
>
> 1 f
This series should be the beginning for bringing some consistency into
the whole bandwidth/io/rate limiting we have all over the place
(vzdump.conf, replication, migration) and allow creating limits for
particular tasks, globally and per-storages, since those are usually
affected by eg. long tasks
Signed-off-by: Wolfgang Bumiller
---
This is the single place where new task types can be added. Feel free to
comment on the ones below.
src/PVE/JSONSchema.pm | 36
1 file changed, 36 insertions(+)
diff --git a/src/PVE/JSONSchema.pm b/src/PVE/JSONSchema.pm
without Sys.Modify on /
(as they can change datacenter.cfg options via the API).
* Storage specific limits apply to users without
Datastore.Allocate access on /storage/X for any involved
storage X.
Signed-off-by: Wolfgang Bumiller
---
PVE/Storage.pm | 84
This will define the global defaults which can be overridden
by per-storage limits.
Signed-off-by: Wolfgang Bumiller
---
data/PVE/Cluster.pm | 1 +
1 file changed, 1 insertion(+)
diff --git a/data/PVE/Cluster.pm b/data/PVE/Cluster.pm
index 5fc7d86..b49fd94 100644
--- a/data/PVE/Cluster.pm
applied whole series
On Fri, Jan 26, 2018 at 11:57:58AM +0100, Dominik Csapak wrote:
> this series adds the ability to show serial ports on the gui,
> and disabling the xtermjs button when no serial device exists for the vm
>
> changes from v1:
> * rebase on current master
> * changed icon to tho
On Wed, Jan 31, 2018 at 12:14:43AM +0100, Geert Stappers wrote:
> Parameter name 'ssh-public-keys' causes trouble
> when written as ssh-public-keys. Because it can read as
> arthmetic expression ssh minus public minus keys.
Just put quotes around it (and/or encode it properly depending on what
you
On Tue, Jan 30, 2018 at 10:34:14AM +0100, Dominik Csapak wrote:
> we want to change the maxIds value even when not rendered, since
> in the wizard, the controller can change from a different panel in the
> tabpanel
>
> Signed-off-by: Dominik Csapak
> ---
> www/manager6/form/ControllerSelector.js
applied both patches
On Thu, Jan 25, 2018 at 01:56:55PM +0100, Thomas Lamprecht wrote:
> Short nack history:
> In PVE 4 Beta we introduced LXC as our new container technology.
> Initially we did not used the our section config format for its
> configuration file in /etc/pve . It was then decided
applied both patches
On Thu, Jan 25, 2018 at 03:55:10PM +0100, Dominik Csapak wrote:
> we always expected MiB/GiB but labeled them incorrectly
>
> Signed-off-by: Dominik Csapak
> ---
> www/manager6/form/DiskStorageSelector.js | 2 +-
> www/manager6/lxc/MPResize.js | 2 +-
> www/mana
without Sys.Modify on /
(as they can change datacenter.cfg options via the API).
* Storage specific limits apply to users without
Datastore.Allocate access on /storage/X for any involved
storage X.
Signed-off-by: Wolfgang Bumiller
---
Changes to v1:
- Paremter order switched
- A limit of 0
applied
On Mon, Jan 29, 2018 at 10:52:39AM +0100, Wolfgang Bumiller wrote:
> We don't use them and newer iproute2 doesn't like ':1' as
> abbreviation for '0:1' apparently.
>
> Signed-off-by: Wolfgang Bumiller
> ---
> src/PVE/Network.pm | 2
applied
On Mon, Jan 29, 2018 at 03:07:48PM +0100, Wolfgang Bumiller wrote:
> There's no technical reason for *us* to limit this to 120.
>
> Signed-off-by: Wolfgang Bumiller
> ---
> src/PVE/JSONSchema.pm | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
&g
applied with v2 of the storage patch
Whether and how more fine grained control will be added is still to be
decided (eg. when the same storage has a different performance on
different cluster nodes).
On Tue, Jan 30, 2018 at 04:34:39PM +0100, Wolfgang Bumiller wrote:
> This series should be
applied
On Thu, Feb 01, 2018 at 02:51:05PM +0100, Wolfgang Link wrote:
> When we clone a VM we also make the smbios unique.
> ---
> [RFC V2]
> Change only uuid and not the hole smbios string.
>
> PVE/QemuServer.pm | 7 +++
> 1 file changed, 7 insertions(+)
>
> diff --git a/PVE/QemuServer.pm
Signed-off-by: Wolfgang Bumiller
---
src/PVE/API2/LXC.pm | 16 +++-
src/PVE/LXC/Create.pm | 12 +++-
2 files changed, 26 insertions(+), 2 deletions(-)
diff --git a/src/PVE/API2/LXC.pm b/src/PVE/API2/LXC.pm
index 733826e..c49cf4f 100644
--- a/src/PVE/API2/LXC.pm
+++ b/src
Signed-off-by: Wolfgang Bumiller
---
This required some refactoring of and currently only handles VMA
archives (wanted to get this out first).
Note that this means we need to extract the config via `vma config ...`
first which of course means this cannot apply to archives coming from a
pipe (but
And add a few helpers for the common cases.
Signed-off-by: Wolfgang Bumiller
---
This has been on my todo list for a while and as an extra
'iface_set_master' is not specific to bridges but can also be used with
eg. vrf devices ;-)
(Note that to test vrf devices you need to set its m
applied
On Wed, Feb 07, 2018 at 10:38:00AM +0100, Thomas Lamprecht wrote:
> ${GITVERSION} is empty else.
>
> Signed-off-by: Thomas Lamprecht
> ---
> Makefile | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/Makefile b/Makefile
> index 550713d..85855ce 100644
> --- a/Mak
applied, thanks
They could have cherry-picked the two function-rename & error handling
commits in between, then they wouldn't have had to resolve the conflicts
they had to resolve ;-) (wouldn't have saved them from auditing the
result though, so I guess it doesn't matter much, but for future
fixup
I already rebased the patches but we didn't get around to doing much
testing yet. You can find the branches on my github if you want to get a
jump on it:
https://github.com/Blub/qemu pve-next (qemu-2.11 pve patch set on top)
https://github.com/Blub/qemu pve-next-patched (above with the extra/ patc
applied the other two of the series, but replaced this one with:
- description => "Whether the drive should be included when making
snapshots.",
+ description => "Controls qemu's snapshot mode feature."
+ . " If activated, changes made to the disk are temporary and will"
+
e the configured limits).
Signed-off-by: Wolfgang Bumiller
---
PVE/Storage.pm| 13
test/run_bwlimit_tests.pl | 75 ++-
2 files changed, 56 insertions(+), 32 deletions(-)
diff --git a/PVE/Storage.pm b/PVE/Storage.pm
index 995ebd3..1
applied
On Fri, Feb 09, 2018 at 08:28:34AM +0100, Thomas Lamprecht wrote:
> allow to access sub elements of configuration items for binding.
>
> E.g., if the following configuration is present:
>
> foo: {
> bar: true
> }
>
> one could do:
>
> xtype: 'displayfield',
> cbind {
> value: '
On Mon, Feb 12, 2018 at 09:27:07AM +0100, Alexandre Derumier wrote:
> +sub round_powerof2 {
Please just use:
return 1 if $_[0] < 2;
return 2 << int(log($_[0]-1)/log(2));
> + local $_ = (shift)-1;
> + my $num = (2**int(log($_)/log(2)))-1;
> + return ++($_ |= $num);
This ties knots into my
Signed-off-by: Wolfgang Bumiller
---
This version avoids the refactoring and double-reading of the VMA file,
thereby supporting reading from a pipe again.
Instead this now uses an extension to vma extract (see the other patch),
and instead of picking the lowest limit altogether, applies limits to
Signed-off-by: Wolfgang Bumiller
---
vma.c | 82 +++
1 file changed, 73 insertions(+), 9 deletions(-)
diff --git a/vma.c b/vma.c
index 04915427c8..91ae14cdc3 100644
--- a/vma.c
+++ b/vma.c
@@ -18,7 +18,9 @@
#include "qemu-com
Checking for the cgroup directory is a kind of time-of-check
time-of-use race condition stop-mode backups seem to
occasionally run into on some systems.
Signed-off-by: Wolfgang Bumiller
---
PVE/QemuServer.pm | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/PVE
applied
On Fri, Feb 16, 2018 at 02:58:10PM +0100, Fabian Grünbichler wrote:
> to avoid an extra dependency
>
> Signed-off-by: Fabian Grünbichler
> ---
> PVE/API2/APT.pm | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/PVE/API2/APT.pm b/PVE/API2/APT.pm
> index 1199c
On Thu, Jul 27, 2017 at 05:06:33PM +0200, Alexandre Derumier wrote:
> Now that proxmox 5.0 is launched, maybe can we try to target cloudinit for
> 5.1 ?
We'd like to move forward with this. I rebased this patch set again
(available at https://github.com/Blub/pve-qemu-server cloudinit branch).
Th
applied, thanks
On Tue, Feb 13, 2018 at 08:51:15AM +0100, Alexandre Derumier wrote:
> qemu 2.11 need a power of 2 cache size.
>
> "
> Parameter 'xbzrle_cache_size' expects is invalid,
> it should be bigger than target page size and a power of two
> "
>
> roundup to near power of 2 value
> ---
>
applied both patches
On Tue, Feb 20, 2018 at 11:10:24AM +0100, Fabian Grünbichler wrote:
> to improve reproducibility of pve-docs, pve-doc-generator and
> pve-doc-mediawiki builds, as well as man pages generated by
> pve-doc-generator when called via dpkg-buildpackage.
>
> timestamps embedded in
Signed-off-by: Wolfgang Bumiller
---
Changes (addressed comments from dominik):
* bwlimit is now an integer, now sure if this is really better, but
sub-kB/s accuracy seems rather useless here
* handle undef when going through storage limits
PVE/API2/Qemu.pm | 11 +++-
PVE/CLI
This time we can't avoid it: nameservers are listed with
separating spaces in LXC and we want to stay consistent and
use the same format in qemu.
---
PVE/QemuServer.pm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index fb36584..8342f87 1
move: don't error out with "you can't move a cdrom"
clone: always full-clone cloud-init images
They get completely replaced anyway at the next start, so
there's no point in keeping them.
Signed-off-by: Wolfgang Bumiller
---
PVE/API2/Qemu.pm | 6 +++--
They're regenerated anyway as needed.
---
PVE/QemuServer.pm | 8 +++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index 95ebb97..319c6a6 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -2057,7 +2057,10 @@ sub vmconfig_undelete_pe
*) always replace old cloudinit images
*) apply pending cloudinit changes when generating a new
image
For cloudinit we now always use vdisk_free before
vdisk_alloc in order to always replace old images, this
allows us to hotplug a new drive by setting it to
`none,media=cdrom` first (to eject the d
Signed-off-by: Wolfgang Bumiller
---
Makefile | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/Makefile b/Makefile
index 1db06ec..4037c3c 100644
--- a/Makefile
+++ b/Makefile
@@ -20,7 +20,7 @@ export PERLDIR=${PREFIX}/share/perl5
PERLINCDIR=${PERLDIR}/asm-x86_64
ARCH
Signed-off-by: Wolfgang Bumiller
---
Used in the cloud-init patch set.
src/PVE/CLIHandler.pm | 14 --
src/PVE/RESTHandler.pm | 38 --
2 files changed, 32 insertions(+), 20 deletions(-)
diff --git a/src/PVE/CLIHandler.pm b/src/PVE/CLIHandler.pm
---
PVE/QemuServer.pm | 12
1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index 172dd5f..95ebb97 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -45,6 +45,8 @@ my $qemu_snap_storage = {rbd => 1, sheepdog => 1};
my $cpu
We now have a patch on top of qemu to allow 'qemu-img dd'
to read from stdin when specifying input and output sizes,
as well as a way to tell it that the size of the source is
not known.
Signed-off-by: Wolfgang Bumiller
---
PVE/QemuServer/Cloudinit.pm | 32 +
From: Alexandre Derumier
Signed-off-by: Alexandre Derumier
Co-developed-by: Wolfgang Bumiller
---
PVE/API2/Qemu.pm| 39 +-
PVE/QemuServer.pm | 124 --
PVE/QemuServer/Cloudinit.pm | 180
PVE
s mac addresses. Ideally
we'd use version 2, but debian stretch ships with a too old
cloud-init for this, so for now we're writing version 1.
Signed-off-by: Wolfgang Bumiller
---
PVE/QemuServer.pm | 6 +
PVE/QemuServer/Cloudinit.pm | 376 -
implement cloudinit
Wolfgang Bumiller (9):
buildsys: use git rev-parse HEAD for GITVERSION
parse_vm_config: allow spaces in values
cloud-init: use qemu-img dd instead of nbd
cloud-init: delete cloudinit images as if they weren't cdroms
cloud-init: don'
This requires a pve-common patch since we need to not only
load the file but also url-encode it.
Signed-off-by: Wolfgang Bumiller
---
PVE/CLI/qm.pm | 15 +++
1 file changed, 15 insertions(+)
diff --git a/PVE/CLI/qm.pm b/PVE/CLI/qm.pm
index d33b949..3347612 100755
--- a/PVE/CLI
applied
On Mon, Mar 05, 2018 at 12:34:03PM +0100, Fabian Grünbichler wrote:
> this patch series adds a 4.13.13-6-pve compatible AA feature set to the
> lxc-pve
> package, replacing the one contained in apparmor >= 2.11.0-3+deb9u2 via
> dpkg-divert. the latter pins AA features to those offered by
On Tue, Mar 06, 2018 at 07:16:32AM +0100, Dietmar Maurer wrote:
>
> > use PVE::Cluster qw (cfs_read_file cfs_write_file);;
> > use PVE::SafeSyslog;
> > @@ -64,7 +65,9 @@ my $check_storage_access = sub {
> >
> > my $volid = $drive->{file};
> >
> > - if (!$volid || $volid eq 'none') {
>
applied
On Fri, Mar 02, 2018 at 12:37:17PM +0100, Dietmar Maurer wrote:
> Changes since V1:
> - split patch into smaller parts
> - new helper $compute_param_mapping_hash()
> - improve parameter names
> - add some inline docs
>
> Besides, functionality should be still the same.
>
>
> Dietmar Mau
applied, will add the same behavior to cloud-init
On Tue, Mar 06, 2018 at 11:08:04AM +0100, Dominik Csapak wrote:
> until now, if either searchdomain or nameservers was not set,
> we used the values from the host.
>
> it would make sense, to set this individually (e.g. only use a
> different sear
Signed-off-by: Wolfgang Bumiller
---
src/PVE/Firewall.pm | 8
1 file changed, 8 insertions(+)
diff --git a/src/PVE/Firewall.pm b/src/PVE/Firewall.pm
index 2feac54..f8d6009 100644
--- a/src/PVE/Firewall.pm
+++ b/src/PVE/Firewall.pm
@@ -1030,6 +1030,7 @@ sub
101 - 200 of 2705 matches
Mail list logo