We hold a lock from snapshot_prepare until snapshot_commit,
so there is no need to copy back the snapshot config to the
actual config.
This allows us to move snapshot_commit to the abstract
common code base.
---
Mirrors changes in QemuServer.pm.
src/PVE/LXC.pm | 6 ++
1 file changed, 2 inse
This sub is not used anywhere.
---
PVE/QemuServer.pm | 25 -
1 file changed, 25 deletions(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index 8defcae..92b0e6a 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -1020,31 +1020,6 @@ sub cleanup_drive_path {
applied
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
applied
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
This enables us to add unused volumes without calling
is_volume_in_use and mimics the behaviour in QemuServer.pm
---
Because is_volume_in_use has different signatures in LXC and QemuServer.pm,
this change allows us to merge more of snapshot_delete into the common code
base.
src/PVE/LXC.pm | 25 ++
Should we then drop the equivalent part in update_pct_config()?
(Only the removal at the end. Its data ($used_volids) is still required
to prevent accidental deletion of volumes.)
On Tue, Mar 01, 2016 at 09:59:11AM +0100, Fabian Grünbichler wrote:
> This enables us to add unused volumes without ca
Replaced ct_modify_file_head_portion() with ct_modify_file()
taking a header and and replacing the section in-place
rather than always moving it to the top of the file.
The behavior for when a section is not defined can be
defined via the provided options (replace, prepend, or
append).
---
src/PV
Changes since v1:
Removed the $header portion so all marks are now simply:
|--- BEGIN PVE ---
|
|--- END PVE ---
The modify function looks for lines matching /---BEGIN PVE.*/ rather
than the complete line since we used 'BEGIN PVE ROUTES' markers in the
RedHat and SUSE container types previously, s
---
src/PVE/LXC/Setup/Base.pm| 101 +++
src/PVE/LXC/Setup/Redhat.pm | 9 +--
src/test/test-centos6-001/etc/hosts.exp | 3 +
src/test/test-debian-001/etc/hosts.exp | 3 +
src/test/test-debian-002/etc/hosts.exp | 3 +
src/test/tes
instead of only when updating the config via pct set or the
API.
This enables us to add unused volumes without calling
is_volume_in_use and mimics the behaviour in QemuServer.pm
---
Changes to v1:
drop similar code from update_pct_config, not needed anymore.
src/PVE/LXC.pm | 36 +
On Fri, Feb 05, 2016 at 10:00:46AM +0100, Dietmar Maurer wrote:
> > > > - my $status = $stat->{status} || 'unknown';
> > > > + my $status = $stat->{qmpstatus} || 'unknown';
> > >
> > > What if the VM is running, but qmp does not work for some reason?
> >
> > vmstatus() ends wi
Hi, qemu devs have send patches to configure backup cluster size:
http://lists.gnu.org/archive/html/qemu-devel/2016-02/msg06062.html
http://lists.gnu.org/archive/html/qemu-devel/2016-02/msg06064.html
Could be great to see if it's improve ceph backup performance.
- Mail original -
De: "
This seems more useful than a 'mounted' state as a
mount-locked container can also be running.
---
src/PVE/CLI/pct.pm | 6 +++---
src/PVE/LXC.pm | 2 ++
2 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/src/PVE/CLI/pct.pm b/src/PVE/CLI/pct.pm
index e599353..6e7fcc0 100755
--- a/sr
> > Ah, OK.
>
> This patch is still missing in the repositories, was there some other
> problem with it, too?
no
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
)
Am 01.03.2016 um 11:03 schrieb Alexandre DERUMIER:
> Hi, qemu devs have send patches to configure backup cluster size:
>
> http://lists.gnu.org/archive/html/qemu-devel/2016-02/msg06062.html
> http://lists.gnu.org/archive/html/qemu-devel/2016-02/msg06064.html
>
>
> Could be great to see if it'
Hi all,
Anybody which have tested the new I/O scheduler blk-mq in the pve4
kernel?
https://www.thomas-krenn.com/en/wiki/Linux_Multi-Queue_Block_IO_Queueing_Mechanism_(blk-mq)
--
Hilsen/Regards
Michael Rasmussen
Get my public GnuPG keys:
michael rasmussen cc
http://pgp.mit.edu:11371/pks/look
applied
> This patch is still missing in the repositories, was there some other
> problem with it, too?
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
applied
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
applied, but I wonder if we can display more useful infos, like "qm list"?
> On March 1, 2016 at 11:07 AM Wolfgang Bumiller wrote:
>
>
> This seems more useful than a 'mounted' state as a
> mount-locked container can also be running.
> ---
> src/PVE/CLI/pct.pm | 6 +++---
> src/PVE/LXC.pm
dh_installinit's -R option uses 'restart' causing a
stop-start cycle with systemd. We really don't want that.
---
debian/postinst | 1 +
debian/rules| 2 +-
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/debian/postinst b/debian/postinst
index 3c6963f..e611f47 100644
--- a/debia
On host level: moved NDP to after connection tracking and
switched to RETURN instead of ACCEPT.
On VM level:
The output direction now uses the accept-mark like the dhcp
option does, too.
Also moved NDP rules below the macfilter & ipset rules.
---
src/PVE/Firewall.pm | 19 ++-
1 fi
We generate ICMP rules from the destination port field,
so allowing them in the source port field only confuses
people.
---
This is a resend of the one already on the list.
src/PVE/Firewall.pm | 33 +
1 file changed, 21 insertions(+), 12 deletions(-)
diff --git a/
compile_iptables_filter() is called twice, once to get the
ipv4 ruleset + ipsets and ones to get the ipv6 ruleset. The
second call still generates ipsets which are discarded so it
makes sense to do this in a separate step.
---
src/PVE/Firewall.pm | 89 ++
When adding an ipset for a device via the 'ipfilter-net$NUM'
name we now implicitly add the default link local address
based on the device's MAC address and a 'nomatch' entry for
the rest of fe80::/10. This is comparable to an ARP/MAC
filter in IPv4 with the main difference that it explicitly
works
get_ipset_cmdlist() had a delete parameter in one commit,
removed in the one after that (dd7a13fddc) and this call
was not updated accordingly with the second patch.
---
src/PVE/Firewall.pm | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/PVE/Firewall.pm b/src/PVE/Firewall.p
This only takes an address and not a CIDR notation. It does
preserve the suffix but ended up compressing
fc00:::/64 to fc00::/64 instead of fc00::/64 and
thus caused the firewall to always show there are pending
changes when ipv6 addresses were available.
---
src/PVE/Firewall.pm | 5 ++
fixes:
* update deprecated field
* move static configuration to class definition
* change applyIf to apply
* change system report handler to pass a function to
check_command (instead of the result of showReport)
Signed-off-by: Dominik Csapak
---
www/manager6/node/Subscription.js | 26 +++
in the option view, change the applyIf to apply to
correctly show the 'Edit' button
Signed-off-by: Dominik Csapak
---
www/manager6/dc/OptionView.js| 2 +-
www/manager6/grid/FirewallOptions.js | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/www/manager6/dc/OptionVie
Signed-off-by: Dominik Csapak
---
www/manager6/node/Config.js | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/www/manager6/node/Config.js b/www/manager6/node/Config.js
index b617c86..b05808f 100644
--- a/www/manager6/node/Config.js
+++ b/www/manager6/node/Config.js
@@ -1
>>Yes it does. We already tested it - it just does not work with VMA. We
>>need to dump qcow2 files cause vma is crashing when the clustersize it
>>get's is not 64kb.
How do you backup to qcow2 ?
@Dietmar : is it possible to improve vma format to handle differents
clustersize(s) ?
- Mail ori
---
www/manager/storage/GlusterFsEdit.js | 8
1 file changed, 8 deletions(-)
diff --git a/www/manager/storage/GlusterFsEdit.js
b/www/manager/storage/GlusterFsEdit.js
index f0d3b4f..fd8f569 100644
--- a/www/manager/storage/GlusterFsEdit.js
+++ b/www/manager/storage/GlusterFsEdit.js
@@ -1
GlusterFS is slower if you use the client and not the QEMU driver.
So to guaranty that GlusterFS perform good allow only KVM/QEMU disk images.
Also remove the vmdk format because it makes no sense. It is much slower then
qcow2.
---
PVE/Storage/GlusterfsPlugin.pm | 4 ++--
1 file changed, 2 insert
May I ask why? What's wrong with keeping your ISO images and/or backups on
gluster?
On Mar 1, 2016, at 3:20 PM, Wolfgang Link wrote:
> GlusterFS is slower if you use the client and not the QEMU driver.
> So to guaranty that GlusterFS perform good allow only KVM/QEMU disk images.
> Also remove t
Because there are situation where you can slow down the KVM/QEMU disk image.
On 03/01/2016 02:22 PM, Stoyan Marinov wrote:
> May I ask why? What's wrong with keeping your ISO images and/or backups on
> gluster?
>
> On Mar 1, 2016, at 3:20 PM, Wolfgang Link wrote:
>
>> GlusterFS is slower if yo
journalctl can check their validness itself
---
src/PVE/Tools.pm | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/src/PVE/Tools.pm b/src/PVE/Tools.pm
index 6303d20..1d1f4b8 100644
--- a/src/PVE/Tools.pm
+++ b/src/PVE/Tools.pm
@@ -1061,7 +1061,7 @@ sub dump_logfile {
}
su
This fixes a timeout which happens when the journal is really big
or/and lies on a slow disk. As journalctl per default reads everything.
As also a single boot may have a duration of over an year lets not use
a boot parameter (as in the previous spin of this series) but rather
the 'since', 'until'
allow the since/until parameters which journalctl understands
We use a really simple regex to check a valit date (time) stamp.
This can be done as this API call gets mainly used by the GUI where
we have full controll over what we send and additional journalctl
has already a good timestamp parser w
This allows the user to only show logs from a specific date span.
For that two date picker and a 'Update' button were added to the
header bar.
For the until date we assume always 23:59:59 o'clock, as witouth
a specific time journalctl assumes 00:00 which would result in an
empty log if we only wan
> Because there are situation where you can slow down the KVM/QEMU disk image.
But it seems to be possible to store CD images currently?
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Yep, storing CD images and backups is currently possible. I personally don’t
think any functionality should be limited because of possibility to slow
something down. This would be like (name vendor here) limiting your ability to
configure a switch port to 10 mbit because it could slow things dow
applied all 3 patches
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
applied
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
applied. But is that update button really necessary?
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
> This only takes an address and not a CIDR notation. It does
> preserve the suffix but ended up compressing
> fc00:::/64 to fc00::/64 instead of fc00::/64 and
> thus caused the firewall to always show there are pending
> changes when ipv6 addresses were available.
So this is a bug in
comment inline:
> diff --git a/src/PVE/Firewall.pm b/src/PVE/Firewall.pm
> index 2e2b2f7..7f8fe98 100644
> --- a/src/PVE/Firewall.pm
> +++ b/src/PVE/Firewall.pm
> @@ -2972,12 +2972,18 @@ sub generate_std_chains {
> }
>
> sub generate_ipset_chains {
> -my ($ipset_ruleset, $clusterfw_conf, $f
On Wed, Mar 02, 2016 at 07:21:33AM +0100, Dietmar Maurer wrote:
> > This only takes an address and not a CIDR notation. It does
> > preserve the suffix but ended up compressing
> > fc00:::/64 to fc00::/64 instead of fc00::/64 and
> > thus caused the firewall to always show there are pen
Am 01.03.2016 um 14:05 schrieb Alexandre DERUMIER:
>>> Yes it does. We already tested it - it just does not work with VMA. We
>>> need to dump qcow2 files cause vma is crashing when the clustersize it
>>> get's is not 64kb.
>
> How do you backup to qcow2 ?
We're currently implementing our own ba
OK, applied all patches.
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
48 matches
Mail list logo