> > > So far, *this all works!*
>
> mmm, and you reassign the same ip address to the new mac ?
>
>>If that is the first free IP in the range, the vNic gets the same IP.
>>Otherwise it will get a new IP.
Can you share your dnsmasq config file ? Maybe do you have extra option
make it working for
Am 08/11/2023 um 13:10 schrieb Aaron Lauterer:
> Ceph does a quick benchmark when creating a new OSD and stores the
> osd_mclock_max_capacity_iops_{ssd,hdd} settings in the config DB.
>
> When destroying the OSD, Ceph does not automatically remove these
> settings. Keeping them can be problematic
since poolid can now contain `/`, it's not possible to use it (properly) as
path parameter anymore.
accordingly:
- merge `read_pool` (`GET /pools/{poolid}`) into 'index' (`GET
/pools/?poolid={poolid}`) (requires clients to extract the only member of the
returned array if they want to query an i
which support nested pools. mostly straight-forward, only pool deletion and the
members grid need some special attention.
Signed-off-by: Fabian Grünbichler
---
Notes:
Thanks Dominik for pointing me at the 'root' part of the proxy :)
www/manager6/dc/PoolView.js | 3 +++
www/manager6/g
Message initial
De: "DERUMIER, Alexandre"
À: pve-devel@lists.proxmox.com ,
aderum...@odiso.com , s.hanre...@proxmox.com
Objet: Re: [pve-devel] [PATCH dnsmasq 0/1] purge previous ip/mac lease
of dhcp reply
Date: 16/11/2023 14:47:20
> > Maybe this [1][2] could be a less intrusive
Am 16/11/2023 um 20:59 schrieb Stoiko Ivanov:
> without this patch the hostname ends up as the empty string in
> run-env-info.json, which results in a parse-error in the TUI code
> (an empty string is not None, but still too short as hostname)
>
> Minimally tested on a VM.
>
> Fixes: bda1cdf699a3
without this patch the hostname ends up as the empty string in
run-env-info.json, which results in a parse-error in the TUI code
(an empty string is not None, but still too short as hostname)
Minimally tested on a VM.
Fixes: bda1cdf699a3fcfc1cf3cfa446b1493689fc8eb8
Signed-off-by: Stoiko Ivanov
-
Am 16/11/2023 um 18:12 schrieb Stoiko Ivanov:
> the recent patch to create /var/lib/vz as dedicated dataset, did so
> for all our products - but this is only needed/wanted for PVE
>
> moved the creation of the root-dataset above the creation of
> rpool/data, so that the pve-specifics can remain in
the recent patch to create /var/lib/vz as dedicated dataset, did so
for all our products - but this is only needed/wanted for PVE
moved the creation of the root-dataset above the creation of
rpool/data, so that the pve-specifics can remain in one if block.
Fixes: dd19d40ceac179ba18652f1d6c3e4c23f
Am 16/11/2023 um 16:00 schrieb Stoiko Ivanov:
> This patch-series contains mostly unrelated patches, which still can be
> tested well together:
> * patch 1/4 fixes the parsing of our kernel cmdline parameters during the
> install (they were ignored if being the last on the line).
> additionally
Parameters needed for booting during installation are best preserved
in the target cmdline as well - e.g. if you need a particular
cmdline switch for your system to boot at all - not having to add it
for the first boot of the installed system and manually adding it to
the bootloader config is an im
v1->v2:
* adapted the creation of /var/lib/vz as separate dataset to be only done
for PVE and not for our other products
original cover-letter for v1:
This patch-series contains mostly unrelated patches, which still can be
tested well together:
* patch 1/4 fixes the parsing of our kernel cmdline
this enables the users to set reservations on / separate from
/var/lib/vz - where backups, ISOs, and other data might fill the
complete pool.
moved the creation of the root-dataset above the creation of
rpool/data, so that the pve-specifics can remain in one if block.
Signed-off-by: Stoiko Ivanov
journald as a core component tries setting a ACL on the journal files
for (non-root) users and fails on our ZFS installs.
Resulting in dmesg being spammed with messages from journald upon each
journal-rotation for each user upon their first login.
This is also suggested by OpenZFS in their Debian
caught me off-guard while debugging other things - the message:
` [: : integer expression expected`
made me look a bit too long at the wrong place.
Fixes: a31259b1597447a0b431cd5c81a6db2bc80f1ddf
Signed-off-by: Stoiko Ivanov
---
unconfigured.sh | 1 +
1 file changed, 1 insertion(+)
diff --git a
this is wrong!
the dataset is only needed for pve
I'll resend a v2
sorry for the noise!
On Thu, 16 Nov 2023 16:00:40 +0100
Stoiko Ivanov wrote:
> this enables the users to set reservations on / separate from
> /var/lib/vz - where backups, ISOs, and other data might fill the
> complete pool.
>
I think you made a mistake when sending this patch.
There are only one 3 out of 4 files attached and
[pve-devel] [PATCH access-control 2/2] pools: record parent/subpool
information
appears to be double.
___
pve-devel mailing list
pve-devel@lists.prox
Tested on a new Proxmox VE 8 cluster. The mclock scheduler settings do
not appear in `ceph config dump` after removing the OSD using the web
UI. Removing a OSD without this setting being set does not cause any
issue either.
Tested-by: Maximiliano Sandoval
Aaron Lauterer writes:
> Ceph does a
and ensure a missing intermediate pool exists at all times.
Signed-off-by: Fabian Grünbichler
---
Notes:
a "missing link" should never happen when modifying via the API (both
deletion
with children and addition without the parent existing is blocked there),
but
it could happen when
and ensure a missing intermediate pool exists at all times.
Signed-off-by: Fabian Grünbichler
---
Notes:
a "missing link" should never happen when modifying via the API (both
deletion
with children and addition without the parent existing is blocked there),
but
it could happen when
with ACLs being inherited along the pool hierarchy.
Signed-off-by: Fabian Grünbichler
---
src/PVE/AccessControl.pm | 10 --
src/test/perm-test6.pl | 16
src/test/test6.cfg | 5 +
3 files changed, 29 insertions(+), 2 deletions(-)
diff --git a/src/PVE/AccessC
this series extends the poolname to allow (for now) two '/' inside the
name to nest pools like this;
parent/child/grandchild
since '/' is a bad character for path parameters, some API shuffling is
needed. other approaches that were tried but discarded, or rejected
outright:
- urlencoding
- system
more or less a copy from the normal section config test, but now with
properties defined multiple times as well as conflicting options
Signed-off-by: Dominik Csapak
---
test/Makefile | 1 +
test/section_config_separated_test.pl | 489 ++
2 files
when using 'init(1)'. This saves the property lists per type instead of
a big one, and using create/updateSchema creates a new schema with the
options as 'oneOf' and/or 'instance-types' (depending if the schemas
match).
with that, we change how we work with the options hash:
it's not needed anymo
a schema can now have the 'oneOf' property which is an array of regular
schemas. In the default case any of that has to match. If the
'type-property'/'instance-types' are given, only the schema for the specific
type will be checked (and handles as 'additionalProperties' if there is
no matching type
to compare nested hashes/lists and scalar values recursively.
Also includes some tests
Signed-off-by: Dominik Csapak
---
changes from v3:
* rename the testfile is_deeply_test to match Makefile
src/PVE/Tools.pm | 31 +
test/Makefile | 1 +
test/is_deeply_test.pl | 142 ++
by simply doing an 'is_deeply' on the generated schema with
the current generated schema
Signed-off-by: Dominik Csapak
---
test/section_config_test.pl | 133
1 file changed, 133 insertions(+)
diff --git a/test/section_config_test.pl b/test/section_config_tes
this series implementes the oneOf schema for the api, see the individual
patches for more details and changelog
only change is the fixed test for pve-common 3/3
(did accidentally include the wrong file when sending the patches)
pve-common:
Dominik Csapak (5):
section config: add test for the s
for parameters only for now, also only implement the basic use case we
want to have currently: use in section config apis where we have more
than one type.
we could improve upon that, e.g. by properly grouping the type relevant
options, and also implementing that for return types.
Signed-off-by:
On 11/16/23 14:55, Dominik Csapak wrote:
> to compare nested hashes/lists and scalar values recursively.
> Also includes some tests
>
> Signed-off-by: Dominik Csapak
> ---
> new in v3, split out from section config
> changes to the is_deeply function:
> * incorporate thomas suggestions
> * fix
caught me off-guard while debugging other things - the message:
` [: : integer expression expected`
made me look a bit too long at the wrong place.
Fixes: a31259b1597447a0b431cd5c81a6db2bc80f1ddf
Signed-off-by: Stoiko Ivanov
---
unconfigured.sh | 1 +
1 file changed, 1 insertion(+)
diff --git a
journald as a core component tries setting a ACL on the journal files
for (non-root) users and fails on our ZFS installs.
Resulting in dmesg being spammed with messages from journald upon each
journal-rotation for each user upon their first login.
This is also suggested by OpenZFS in their Debian
This patch-series contains mostly unrelated patches, which still can be
tested well together:
* patch 1/4 fixes the parsing of our kernel cmdline parameters during the
install (they were ignored if being the last on the line).
additionally everything not used by our installer now end up in the
Parameters needed for booting during installation are best preserved
in the target cmdline as well - e.g. if you need a particular
cmdline switch for your system to boot at all - not having to add it
for the first boot of the installed system and manually adding it to
the bootloader config is an im
this enables the users to set reservations on / separate from
/var/lib/vz - where backups, ISOs, and other data might fill the
complete pool.
Signed-off-by: Stoiko Ivanov
---
Proxmox/Install.pm | 3 +++
1 file changed, 3 insertions(+)
diff --git a/Proxmox/Install.pm b/Proxmox/Install.pm
index c
basic CRUD for the profile section config
Signed-off-by: Dominik Csapak
---
changes from v2:
* add type parameter so we can filter the list
(useful for the ui)
PVE/API2/Cluster.pm | 7 +
PVE/API2/Cluster/Makefile| 1 +
PVE/API2/Cluster/Profiles.pm | 239 ++
we have to that here, so the properties/options are correctly configured
when using that feature on the cli
Signed-off-by: Dominik Csapak
---
changes from v2:
* call init() instead of init(1) since the plugin already does that for us
src/PVE/CLI/pct.pm | 6 ++
1 file changed, 6 insertions(+)
we have to that here, so the properties/options are correctly configured
when using that feature on the cli
Signed-off-by: Dominik Csapak
---
changes from v2:
* call init() instead of init(1) since the plugin does that already for us
PVE/CLI/qm.pm | 6 ++
1 file changed, 6 insertions(+)
dif
we use the the profile cfg as the 'param' hash, but overwrite the values
with the ones from the api call, so one can overwrite options from the
profile easily
also we add the used profile to the meta info in the config, since
it might be interesting which one was used
Signed-off-by: Dominik Csapa
This series aims to provide profile support when creating guests (ct/vm)
so that users can reuse options without having to specify them every
time.
UI is still to follow, depends on [0] for the oneOf schema support.
changes from v2:
* rename get_guest_ready_config into load_profile
* rename some
"DERUMIER, Alexandre" writes:
>>>So far, *this all works!*
>
> mmm, and you reassign the same ip address to the new mac ?
>
If that is the first free IP in the range, the vNic gets the same IP.
Otherwise it will get a new IP.
___
pve-devel mailing li
we use the profile cfg as the 'param' hash, but overwrite the values
with the ones from the api call, so one can overwrite options from
the profile easily
Signed-off-by: Dominik Csapak
---
changes from v2:
* adapt to load_profile name change
src/PVE/API2/LXC.pm | 23 +++
1 f
simply uses the json_config_properties for the ct config and maps them
to "ct_${opt}"
Signed-off-by: Dominik Csapak
---
src/PVE/Makefile | 1 +
src/PVE/Profiles/CT.pm| 28
src/PVE/Profiles/Makefile | 4
3 files changed, 33 insertions(+)
create mo
this is intended to house custom profiles which can be used
on guest creation instead of manually needing to specify every option.
we do special things here:
* we always set 'allow_unknown' to 1, because when using the guest
specific parts in the cli, we cannot depend on the other one, else
we
simply uses the json_config_properties for the vm config and maps them
to "vm_${opt}"
Signed-off-by: Dominik Csapak
---
PVE/Makefile | 1 +
PVE/Profiles/Makefile | 5 +
PVE/Profiles/VM.pm| 28
3 files changed, 34 insertions(+)
create mode 100644 P
Signed-off-by: Dominik Csapak
---
src/PVE/Cluster.pm | 1 +
src/pmxcfs/status.c | 1 +
2 files changed, 2 insertions(+)
diff --git a/src/PVE/Cluster.pm b/src/PVE/Cluster.pm
index cfa2583..c01bf89 100644
--- a/src/PVE/Cluster.pm
+++ b/src/PVE/Cluster.pm
@@ -80,6 +80,7 @@ my $observed = {
's
more or less a copy from the normal section config test, but now with
properties defined multiple times as well as conflicting options
Signed-off-by: Dominik Csapak
---
test/Makefile | 1 +
test/section_config_separated_test.pl | 489 ++
2 files
when using 'init(1)'. This saves the property lists per type instead of
a big one, and using create/updateSchema creates a new schema with the
options as 'oneOf' and/or 'instance-types' (depending if the schemas
match).
with that, we change how we work with the options hash:
it's not needed anymo
a schema can now have the 'oneOf' property which is an array of regular
schemas. In the default case any of that has to match. If the
'type-property'/'instance-types' are given, only the schema for the specific
type will be checked (and handles as 'additionalProperties' if there is
no matching type
for parameters only for now, also only implement the basic use case we
want to have currently: use in section config apis where we have more
than one type.
we could improve upon that, e.g. by properly grouping the type relevant
options, and also implementing that for return types.
Signed-off-by:
this series implementes the oneOf schema for the api, see the individual
patches for more details and changelog
pve-common:
Dominik Csapak (5):
section config: add test for the schemas
tools: add is_deeply
json schema: implement 'oneOf' schema
section config: allow separated property list
by simply doing an 'is_deeply' on the generated schema with
the current generated schema
Signed-off-by: Dominik Csapak
---
test/section_config_test.pl | 133
1 file changed, 133 insertions(+)
diff --git a/test/section_config_test.pl b/test/section_config_tes
to compare nested hashes/lists and scalar values recursively.
Also includes some tests
Signed-off-by: Dominik Csapak
---
new in v3, split out from section config
changes to the is_deeply function:
* incorporate thomas suggestions
* fix sub call style
* add comment to make it clear `ref` never ret
the debug log without my patch:
vm with 12:45:db:3a:04:97 got 192.168.2.10
Nov 15 21:33:31 formationkvm3 dnsmasq-dhcp[796025]: DHCPDISCOVER(vnetpve)
192.168.2.10 12:45:db:3a:04:97
Nov 15 21:33:31 formationkvm3 dnsmasq-dhcp[796025]: DHCPOFFER(vnetpve)
192.168.2.10 12:45:db:3a:04:97
Nov 15 21:3
>>Maybe this [1][2] could be a less intrusive solution for this issue?
Yes, dhcp release packet should be the way, but I don't known if can
simply forge packet why any mac ?
I'll test it this afternoon to see if it's work.
___
pve-devel mailing list
pve
>>I experimented with several approaches with dnsmasq leases.
>>I cannot reproduce your example because it works in my examples.
>>My procedure:
>>dnsmasq config:
dhcp-range=set:DHCPNAT-10.1.0.0-16,10.1.0.0,static,255.255.0.0,infinite
>>VM 108 net1: MAC: bc:24:11:ad:0e:2e
>>
>>qm set 108 --dele
On Thu, Nov 16, 2023 at 12:50:44PM +0100, Filip Schauer wrote:
> Add a dev[n] argument to the container config to pass devices through to
> a container. A device can be passed by its path. Additionally the access
> mode, uid and gid can be specified through their respective properties.
>
> Signed-
The latest change made the MACs monospaced, hence they do require more
space now. A multiple of 17 was chosen to match the number of
characters.
Signed-off-by: Maximiliano Sandoval R
---
www/manager6/qemu/AgentIPView.js | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/www/mana
It becomes easier to read when you have multiple rows showing MACs when
they have the same size.
Signed-off-by: Maximiliano Sandoval R
---
www/manager6/qemu/AgentIPView.js | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/www/manager6/qemu/AgentIPView.js b/www/manager6/qe
"DERUMIER, Alexandre" writes:
>
> I can reproduce easily 100%:
>
>
> - create a nic with mac:xx:xx:xx:xY ip: 192.168.0.10
> - start vm. (the ether file is filed + reload)
> - the guest do a dhcp request, the dnsmasq respond a store the lease
> in /var/lib/misc/zone.lease
>
> - delete the nic
>
Signed-off-by: Dominik Csapak
---
src/window/NotificationMatcherEdit.js | 84 ++-
1 file changed, 81 insertions(+), 3 deletions(-)
diff --git a/src/window/NotificationMatcherEdit.js
b/src/window/NotificationMatcherEdit.js
index c0dfa34..0f29203 100644
--- a/src/window/No
here a few things i could fix, sorry for being so late with this
there are still a few non optimal things in the ux, but nothing i could
fix in a quick manner
not that my fixes are just to make it work properly and the most
important UX issues. There are still some things that can be improved
(UX
by removing the 'invert' checkbox and instead show the 4 modes possible,
we still assemble/parse the invert for the backend
Signed-off-by: Dominik Csapak
---
src/window/NotificationMatcherEdit.js | 54 +--
1 file changed, 18 insertions(+), 36 deletions(-)
diff --git a/sr
'delete' is only possible for editing jobs, not creating them
Signed-off-by: Dominik Csapak
---
follow up to lukas patches
www/manager6/dc/Backup.js | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/www/manager6/dc/Backup.js b/www/manager6/dc/Backup.js
index e1c76a1d..9a
Patch v5 available
https://lists.proxmox.com/pipermail/pve-devel/2023-November/060285.html
On 16/11/2023 09:32, Wolfgang Bumiller wrote:
On Wed, Nov 15, 2023 at 03:14:50PM +0100, Thomas Lamprecht wrote:
concept wise this looks pretty much OK, but a few (mostly code-style) comments
in line
Am
Add a dev[n] argument to the container config to pass devices through to
a container. A device can be passed by its path. Additionally the access
mode, uid and gid can be specified through their respective properties.
Signed-off-by: Filip Schauer
---
Changes since v4:
* Rename device lists to "mo
Am 16.11.23 um 11:45 schrieb Philipp Hufnagl:
>
> I was trying to order the commits in a way that, while they are
> changing the code, the should not impact the behavior of the program.
> It should be possible to apply all previous commits without modifying
> the behavior.
>
> If it is easier or
On 11/16/23 10:57, Fiona Ebner wrote:
> Am 15.11.23 um 17:00 schrieb Philipp Hufnagl:
>> @@ -831,10 +855,27 @@ sub check_datastore_exists {
>> return 1;
>> }
>> }
>> -
>> die "$storeid: Cannot find datastore '$datastore', check permissions
>> and existence!\n";
>> }
>>
Am 10.11.23 um 10:33 schrieb Hannes Duerr:
> adds vendor and product information for SCSI devices to the json schema and
> checks in the VM create/update API call if it is possible to add these to
> QEMU as a device option
>
> Signed-off-by: Hannes Duerr
> ---
> PVE/API2/Qemu.pm| 12 +++
This adds a check if the datastore and the namespace is valid when a
user attempts to add a new PBS datastore.
Since the namespace only can be checked after the datastore is
validated, the datastore will be checked as well, regardless that it
will be done later in the superclass anyway.
The funct
Because it is needed later in this patch series, the method
pbs_api_connect is moved earlyer in the code
Signed-off-by: Philipp Hufnagl
---
src/PVE/Storage/PBSPlugin.pm | 63 ++--
1 file changed, 32 insertions(+), 31 deletions(-)
diff --git a/src/PVE/Storage/PBSP
Parts contained in activate_storage are needed to be run to fix #5008,
however, implementing a namespace check there would cause unneeded
overhead.
Therefore, this patch extracts the method check_datastore_exists from
activate storage.
Signed-off-by: Philipp Hufnagl
---
src/PVE/Storage/PBSPlugi
Currently, when adding a PBS storage with a namespace that does not
exist, the storage gets added normally, but browsing/using it only
returns a cryptic error message.
This change checks if the namespace entered when adding is valid and
prompts an error if it is not. If no namespace is provided, t
It would be nice to reuse an existing PBS connection for scan_datastore.
Because scan_datastore is used multiple in the code, it can not be
changed without breaking existing code.
This change add an optional connection parameter to scan_datastore. If
it is passed it will use this connection. If no
Nit: Like can be seen from the "and" in the commit title, it's actually
two changes. Moving the dependencies first, and then introducing the new
helper in a second patch would be even better. But no big deal :)
Am 10.11.23 um 10:33 schrieb Hannes Duerr:
> +sub get_scsi_devicetype {
> +my ($dri
Am 15.11.23 um 17:00 schrieb Philipp Hufnagl:
> @@ -831,10 +855,27 @@ sub check_datastore_exists {
> return 1;
> }
> }
> -
> die "$storeid: Cannot find datastore '$datastore', check permissions and
> existence!\n";
> }
>
> +sub check_namespace_exists {
> +my ($clas
> On 15.11.2023 17:00 CET Philipp Hufnagl wrote:
>
>
> This adds a check if the datastore and the namespace is valid when a
> user attempts to add a new PBS datastore.
>
> Since the namespace only can be checked after the datastore is
> validated, the datastore will be checked as well, regar
Typi in heading s/Extraxt/Extract, maybe call it factor out instead.
> On 15.11.2023 17:00 CET Philipp Hufnagl wrote:
>
>
> Parts contained in activate_storage are needed to be run to fix #5008,
> however, implementing a namespace check there would cause unneded
> overhead.
Typo s/unneded/unn
Maybe this [1][2] could be a less intrusive solution for this issue?
[1] https://manpages.ubuntu.com/manpages/focal/en/man1/dhcp_release.1.html
[2] https://packages.debian.org/de/sid/dnsmasq-utils
___
pve-devel mailing list
pve-devel@lists.proxmox.com
> On 15.11.2023 17:00 CET Philipp Hufnagl wrote:
>
>
> It would be nice to reuse an existing PBS connection for scan_datastore.
> Because scan_datastore is used multiple in the code, it can not be
> canged without breaking existing code.
Typo s/canged/changed.
>
> This change add an optio
On Wed, Nov 15, 2023 at 03:14:50PM +0100, Thomas Lamprecht wrote:
> concept wise this looks pretty much OK, but a few (mostly code-style)
> comments in line
>
> Am 13/11/2023 um 11:30 schrieb Filip Schauer:
> > diff --git a/src/PVE/LXC/Config.pm b/src/PVE/LXC/Config.pm
> > index 56e1f10..9f325f2
81 matches
Mail list logo
