With Qemu 4.2 we encountered a problem with unix sockets and SSH socket
forwarding for drive-mirror. It seems the socket gets reopened again and
again after it closes for some reason. This can be worked around by
specifying 'block-job-cancel' instead of 'block-job-complete' when we're
not
As the NBD server spawned by qemu can only listen on a single socket,
we're dependent on a version being passed to vm_start that indicates
which protocol can be used, TCP or Unix, by the source node.
The change in socket type (TCP to Unix) comes with a different URI. For
unix sockets it has the
The reuse of the tunnel, which we're opening to communicate with the target
node and to forward the unix socket for the state migration, for the NBD unix
socket requires adding support for an array of sockets to forward, not just a
single one. We also have to change the $sock_addr variable to an
For secure live migration with local disks via NBD over a unix socket,
we have to somehow communicate from the source node to the target node
if it supports it. This is because there can only be one NBD server with
exactly one socket bound.
The source node passes that information via STDIN.
Currently NBD storage migration always uses unencrypted TCP. The
following 4 patches add support for unix sockets that are forwarded over
SSH.
For backwards compatibility this requires some kind of communication
from the source node to the target node, because the NBD server can only
be started
On March 16, 2020 4:44 pm, Aaron Lauterer wrote:
> Move the logic which volumes are included in the backup job to its own
> method and adapt the VZDump code accordingly. This makes it possible to
> develop other features around backup jobs.
>
> Signed-off-by: Aaron Lauterer
> ---
>
> v2 -> v3:
On March 16, 2020 4:44 pm, Aaron Lauterer wrote:
> This extracts the logic which guests are to be included in a backup job
> into its own method 'get_included_guests'. This makes it possible to
> develop other features around backup jobs.
>
> Logic which was spread out accross the API2/VZDump.pm
On 3/17/20 2:40 PM, Thomas Lamprecht wrote:
On 3/17/20 2:33 PM, Wolfgang Bumiller wrote:
On 3/17/20 2:24 PM, Thomas Lamprecht wrote:
On 3/17/20 2:10 PM, Wolfgang Bumiller wrote:
On 3/17/20 12:31 PM, Thomas Lamprecht wrote:
On 3/17/20 10:27 AM, Wolfgang Bumiller wrote:
On 3/17/20 7:35 AM,
On 3/17/20 2:33 PM, Wolfgang Bumiller wrote:
> On 3/17/20 2:24 PM, Thomas Lamprecht wrote:
>> On 3/17/20 2:10 PM, Wolfgang Bumiller wrote:
>>> On 3/17/20 12:31 PM, Thomas Lamprecht wrote:
On 3/17/20 10:27 AM, Wolfgang Bumiller wrote:
> On 3/17/20 7:35 AM, Thomas Lamprecht wrote:
>>
On 3/17/20 2:24 PM, Thomas Lamprecht wrote:
On 3/17/20 2:10 PM, Wolfgang Bumiller wrote:
On 3/17/20 12:31 PM, Thomas Lamprecht wrote:
On 3/17/20 10:27 AM, Wolfgang Bumiller wrote:
On 3/17/20 7:35 AM, Thomas Lamprecht wrote:
CONTAINER_INTERFACE[0] is omething systemd people call their API and
On 3/17/20 2:10 PM, Wolfgang Bumiller wrote:
> On 3/17/20 12:31 PM, Thomas Lamprecht wrote:
>> On 3/17/20 10:27 AM, Wolfgang Bumiller wrote:
>>> On 3/17/20 7:35 AM, Thomas Lamprecht wrote:
CONTAINER_INTERFACE[0] is omething systemd people call their API and
we need to adapt to it a bit,
On 3/17/20 12:31 PM, Thomas Lamprecht wrote:
On 3/17/20 10:27 AM, Wolfgang Bumiller wrote:
On 3/17/20 7:35 AM, Thomas Lamprecht wrote:
CONTAINER_INTERFACE[0] is omething systemd people call their API and
we need to adapt to it a bit, even if it means doing stupid
unnecessary things, as else
Improvement of grammar and punctuation.
Clarify the HA limitations.
Remove future tense in some sentences.
It is not good to use it in technical/scientific papers.
Rewrite some sentences to improve understanding.
---
pvesr.adoc | 112 ++---
1 file
On March 17, 2020 12:40 pm, Thomas Lamprecht wrote:
> On 3/17/20 11:21 AM, Stefan Reiter wrote:
>>> +$local_volumes->{$opt} = $conf->{${opt}};
>>
>> Does $conf->{${opt}} have too many brackets or is this another arcane perl
>> syntax I've yet to discover? (iow. why not just
Moved code so that initialization of drivedesc_hash stays a single block.
Avoid auto-vivication in parse_drive.
Signed-off-by: Fabian Ebner
---
Follow-up suggested by Fabian G.
I checked the call-sites for is_valid_drivename and in most cases
it won't benefit from including unused drives,
On 17/03/2020 12:40, Thomas Lamprecht wrote:
On 3/17/20 11:21 AM, Stefan Reiter wrote:
+$local_volumes->{$opt} = $conf->{${opt}};
Does $conf->{${opt}} have too many brackets or is this another arcane perl syntax
I've yet to discover? (iow. why not just $conf->{$opt} ?)
It's not
On 3/17/20 11:21 AM, Stefan Reiter wrote:
>> +$local_volumes->{$opt} = $conf->{${opt}};
>
> Does $conf->{${opt}} have too many brackets or is this another arcane perl
> syntax I've yet to discover? (iow. why not just $conf->{$opt} ?)
It's not that arcane, you surely used it
On 3/17/20 10:27 AM, Wolfgang Bumiller wrote:
> On 3/17/20 7:35 AM, Thomas Lamprecht wrote:
>> CONTAINER_INTERFACE[0] is omething systemd people call their API and
>> we need to adapt to it a bit, even if it means doing stupid
>> unnecessary things, as else systemd decides to regress and suddenly
On 3/17/20 11:48 AM, Wolfgang Bumiller wrote:
> On 3/17/20 7:35 AM, Thomas Lamprecht wrote:
>> Else some newer system do not see the interface as up and refuse to
>> manage it..
>
> Where do you run into that actually? I can get my arch container to behave
> normally with just adding
Since it's been a while, here's a rebased version the two still missing patches
of my previous RFC[0].
Except for a slight rewording of the message in patch 2 it's exactly the same,
so consider this a RESEND more than anything. I think the series might have
gotten lost a bit after parts where
Instead of the old 'ring_addr' property (which is kept for
compatibility), we also encode the link numbers into the new peerLinks
structure. This allows us to display which IP is assigned to which link
on the cluster in the join dialog, helping a user identify which link
should receive which
CorosyncLinkEdit is a Panel that contains between one and 8
CorosyncLinkSelectors. These can be added or removed with according
buttons.
Values submitted to the API are calculated by each
ProxmoxNetworkSelector itself. This works because ExtJS searches
recursively through all child components for
On 3/17/20 7:35 AM, Thomas Lamprecht wrote:
Else some newer system do not see the interface as up and refuse to
manage it..
Where do you run into that actually? I can get my arch container to
behave normally with just adding `lxc.mount.auto = sys:mixed` to
/etc/pve/lxc/$vmid.conf atm.
Casually looked over the patches, looking good to me so far - I'll give
them some testing later.
Two things inline.
On 17/03/2020 08:55, Fabian Grünbichler wrote:
with incremental drive-mirror and dirty-bitmap tracking.
1.) get replicated disks that are currently referenced by running VM
2.)
On 3/17/20 10:27 AM, Wolfgang Bumiller wrote:
On 3/17/20 7:35 AM, Thomas Lamprecht wrote:
CONTAINER_INTERFACE[0] is omething systemd people call their API and
we need to adapt to it a bit, even if it means doing stupid
unnecessary things, as else systemd decides to regress and suddenly
break
On 3/17/20 7:35 AM, Thomas Lamprecht wrote:
CONTAINER_INTERFACE[0] is omething systemd people call their API and
we need to adapt to it a bit, even if it means doing stupid
unnecessary things, as else systemd decides to regress and suddenly
break network stack in CT after an upgrade[1].
This
by re-using a dirty bitmap that represents changes since the divergence
of source and target volume. requires a qemu that supports incremental
drive-mirroring, and will die otherwise.
Signed-off-by: Fabian Grünbichler
---
Notes:
v1-v2:
- use newer Qemu patches picked up by me and
I recently picked up and finished some work-in-progress patches for
adding bitmap support to drive-mirror (it got added to backup block jobs
in 4.0, with plenty of fixes in 4.1 and 4.2) and submitted them
upstream.
IMHO this is in a shape now where we can include it, but I'd also be
fine with
to make migration logs a bit easier to grasp with a quick glance.
Signed-off-by: Fabian Grünbichler
---
Notes:
unchanged since v1
PVE/QemuMigrate.pm | 11 +++
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/PVE/QemuMigrate.pm b/PVE/QemuMigrate.pm
index
with incremental drive-mirror and dirty-bitmap tracking.
1.) get replicated disks that are currently referenced by running VM
2.) add a block-dirty-bitmap to each of them
3.) replicate ALL replicated disks
4.) pass bitmaps from 2) to drive-mirror for disks from 1)
5.) skip replicated disks when
Hi,
On 3/16/20 2:34 PM, Roland Kammerer wrote:
> Hi all,
>
> Commit[1] extended the filter, but if you look closely, there is no ','
> between the second and third rule. Quite frankly, I don't know what LVM
> makes out of that, most likely it even considers them as 2 separate
> rules. Or it
Else some newer system do not see the interface as up and refuse to
manage it..
Signed-off-by: Thomas Lamprecht
---
src/PVE/LXC.pm | 1 +
1 file changed, 1 insertion(+)
diff --git a/src/PVE/LXC.pm b/src/PVE/LXC.pm
index 0742a53..f811550 100644
--- a/src/PVE/LXC.pm
+++ b/src/PVE/LXC.pm
@@
CONTAINER_INTERFACE[0] is omething systemd people call their API and
we need to adapt to it a bit, even if it means doing stupid
unnecessary things, as else systemd decides to regress and suddenly
break network stack in CT after an upgrade[1].
This mounts the parent /sys as ro, child mounts can
33 matches
Mail list logo