Thomas Lamprecht [Fri, 19 Jul 2019 13:42:13 +0000 (15:42 +0200)]
debian: bump compat to 12 and don't restart container.slice
since compat 10 the restart is default, as I want to use
'dh_installsystemd' (vs 'dh_systemd_start') I need at least compat
level 11, so go for the now recommended compat level 12.
diffoscope tells me that the main change us the wanted:
./postinst
> @@ -1,10 +1,15 @@
> #!/bin/sh
> set -e
> -# Automatically added by dh_systemd_start/12.1.1
> +# Automatically added by dh_installsystemd/12.1.1
> if [ "$1" = "configure" ] || [ "$1" = "abort-upgrade" ] || [ "$1" = "abort-deconfigure" ] || [ "$1" = "abort-remove" ] ; then
> if [ -d /run/systemd/system ]; then
> systemctl --system daemon-reload >/dev/null || true
> - if [ -n "$2" ]; then
> - _dh_action=restart
> - else
> - _dh_action=start
> - fi
> - deb-systemd-invoke $_dh_action 'system-pve\x2dcontainer.slice' >/dev/null || true
> + deb-systemd-invoke start 'system-pve\x2dcontainer.slice' >/dev/null || true
> fi
> fi
> # End automatically added section
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Thomas Lamprecht [Wed, 17 Jul 2019 10:07:40 +0000 (12:07 +0200)]
setup getty: ensure the getty.target is not masked
some distro templates have this masked by default, it makes sense to
always ensure that it can work, a CT admin can still prevent this by
using the .pve-ignore.$file mechanism.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com> Acked-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Thomas Lamprecht [Thu, 18 Jul 2019 15:17:17 +0000 (17:17 +0200)]
setup getty: drop now obsolete setup_systemd_console
The setup_container_getty_service can now handle also old
getty@.service if the newer container-getty@.service is not
available. So drop, and convert the two remaining users to calling
the now compatible setup_container_getty_service
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com> Acked-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Some recent distributions running as a LXC container eat the relative
low default limits up very fast. Thus increase all those
(semi-related) limits by a factor of 512. This was chosen by using
one of our bigger know CT setup (~1500 CTs per host) and the fact
that I can have only a very low count (circa 5 - 7) of running
"inotify watch hungry" CTs (e.g., ones with a recent systemd > 240).
So, as 5 * 512 is well >> 1500, we can assume with confidence to
allow most reasonable and existing setups by default.
As with the kernel commit d46eb14b735b11927d4bdc2d1854c311af19de6d
"fs: fsnotify: account fsnotify metadata to kmemcg" [0] the memory
usage from the watch and queue overhead is accounted to the users
respective memory CGroup (i.e., for LXC containers their memory
limit) we can do this without to much fear of negative implications.
Don't change the hardcoded kernel default values directly though,
ship a sysctl.d configuration file, which is a bit more transparent
about what happens and can be shipped by the component needing this
(i.e., pve-container).
Follow the considerations of `man 5 sysctl.d` for shipping:
> Packages should install their configuration files in /lib/. Files
> in /etc/ are reserved for the local administrator, who may use this
> logic to override the configuration files installed by vendor
> packages. All configuration files are sorted by their filename in
> lexicographic order, regardless of which of the directories they
> reside in. If multiple files specify the same option, the entry in
> the file with the lexicographically latest name will take
> precedence. It is recommended to prefix all filenames with a
> two-digit number and a dash, to simplify the ordering of the files.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Stefan Reiter [Tue, 9 Jul 2019 15:20:57 +0000 (17:20 +0200)]
fix #2270: allow custom lxc options to be restored as root
Seems to be a regression introduced with f360d7f16b094fa258cf82d2557d06f3284435e4 (related to #2028).
$conf->{'lxc'} would always be defined, hence we never replaced it with
the restored options.
Co-developed-by: Oguz Bektas <o.bektas@proxmox.com> Signed-off-by: Stefan Reiter <s.reiter@proxmox.com>
Oguz Bektas [Fri, 5 Jul 2019 11:27:05 +0000 (13:27 +0200)]
fix #1451: allow one to add mount options to CT mountpoints
for now allows the following non-problematic ones:
* noexec - Do not permit execution of binaries on the mounted FS
* noatime - Do not update inode access times on this filesystem
* nosuid - Do not allow suid or sgid bits to take effect
* nodev - Do not interpret character or block devices on the FS
Dominic Jäger [Wed, 12 Jun 2019 10:04:57 +0000 (12:04 +0200)]
Fix #576: Fix dangling files for Move Disk
When Move Disk is called for a container rsync starts copying it to a
new destination. This initial rsync process gets killed when the Stop
button gets pressed. At this moment the destination file is not fully
copied and useless as a consequence. Our code already tries to remove
it. However, rsync has forked and those forks are still accessing the
destination file for some time. Thus, the attempt to remove it fails.
With the patch we wait for other processes to release the destination
files. As we are in a mount namespace and protected by a config lock,
those other processes should be children of rsync only. The waiting
time was less than a second when I tried it. Afterwards, the existing
remove procedure is carried out.
Co-developed-by: Wolfgang Bumiller <w.bumiller@proxmox.com> Signed-off-by: Dominic Jäger <d.jaeger@proxmox.com>
Alwin Antreich [Thu, 23 May 2019 07:13:40 +0000 (09:13 +0200)]
Fix: check if compression_map format is undefined
We want to check for an supported compression type, but the check was
not correct as this only works if both sides are scalars, but an
assignment to an array is always "truthy", so actually check explicitly
if the compression type is supported before.
Signed-off-by: Alwin Antreich <a.antreich@proxmox.com> Co-authored-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
This allows to have the same semantics as qemu-server:
* immediate hard-kill
* shutdown with kill after timeout
* shutdown without kill after timeout
And thus we finally can move the vm_shutdown API call to a correct
semantic, i.e., do not immediate hard kill if forceStop is not passed
but rather see it as stop after timeout knob.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
No call-site used this parameter, and thus it was dead code,
remove it not only for cleanup sake but also to make space for a new
"nokill-after-timeout" parameter, comming in a future patch.
This code was always dead since it was introduced with the addition
of vm_stop in commit b1bad293c4f7a6024bbd363b6784b3875ca5d098
so pretty safe to remove anyway.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Stoiko Ivanov [Mon, 6 May 2019 14:27:44 +0000 (16:27 +0200)]
raise supported fedora version to 30
Tested by installing a fedora 29 container and upgrading it via dnf [0].
The upgraded container boots, but in order to get networking running (and many
warnings and errors less in the journal) 'nesting' needs to be activated both
for privileged and unprivileged containers.
Christian Ebner [Wed, 17 Apr 2019 14:38:28 +0000 (16:38 +0200)]
fix: #1075: Correctly restore CT templates form backup
Restoring a backup from a CT template wrongly resulted in a CT with the template
flag set in the config.
This makes sure the CT template backup gets restored to a CT and only if the
storage supports templates, the resulting CT is converted to a template.
Otherwise the backup restores simply to a CT.
Signed-off-by: Christian Ebner <c.ebner@proxmox.com>
Unconditionally add a '--bwlimit' parameter to the rsync invocation, defaulting
to an argument of '0' (= unlimited - see `man rsync).
Normally this is a rate per second, with a passed unit. With no unit
passed rsync assumes "K", which is exactly what our units are in, so
make our life easy and omit it.
Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com> Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Oguz Bektas [Fri, 29 Mar 2019 16:16:33 +0000 (17:16 +0100)]
fix #2147: re-add support for current opensuse tumbleweed
this enables opensuse-tumbleweed templates to be used in Proxmox VE
_again_. It was already supported but it seems that the os-release
backed ID changed and thus our distro detection code didn't detect it
anymore.
a few things didn't work properly in my tests, so some things to consider:
* (probably) because of network configuration issues, it takes a while
for the container to start fully (~30s on my setup)
* unprivileged containers (w/ and w/o nesting enabled) had no network
after starting, and needed to be enabled manually with ip addr and
route.
* privileged containers seemed to function normally, except the
startup delay
Oguz Bektas [Thu, 28 Mar 2019 13:01:44 +0000 (14:01 +0100)]
fix #1607: implement pct fstrim
runs fstrim on the rootfs and all mountpoints of a given container. this
works for both running and stopped containers.
lock the CT during this operation using a config lock as it is
potentially long running. While fstrim itself wouldn't really need
the lock, as multiple parallel fstrim calls can be made without
problems, we want to forbid migrations during it and want to avoid
that we unmount a with the CT mounted with 'mount' lock (race) -
while we could handle and allow this its just not needed and easier
this way
Signed-off-by: Oguz Bektas <o.bektas@proxmox.com> Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
Stoiko Ivanov [Tue, 19 Mar 2019 15:34:27 +0000 (16:34 +0100)]
setup: fix alpine ipv6-slaac configuration
busybox ifupdown implementation differs from debian's - configuration type
auto is not supported. If SLAAC is selected for the ipv6 configuration of an
interface, the complete networking is not started, because of that error.
This workaround sets the interface type to 'manual' in case SLAAC is selected
(as is already done for dhcpv6 (for different reasons)). That way all other
configuration stanzas are setup correctly, and if a ipv4 configuration is
present for the same interface the SLAAC-part usually works out of the box
anyways (unless 'accept_ra' is set to 0 for the interface in the kernel).
Alwin Antreich [Fri, 8 Mar 2019 14:41:55 +0000 (15:41 +0100)]
Fix #2109: resize rbd volume for container failed
On resizing a container's disk image the filesystem is extended and in
the case of RBD the returned path of the volume was not a path to a
mapped device.
This patch uses map_volume (respectively unmap_volume) to get a device
mapped and its path returned by the storage plugin. If a path is not
returned then the path method is tried. Currently only the RBD storage
plugin returns a path on map_volume.
Oguz Bektas [Mon, 4 Mar 2019 10:02:53 +0000 (11:02 +0100)]
fix #2117: don't keep custom idmap in pct pipe restore
while doing a pct restore operation, custom id mappings were being
obtained from the archive file to be used in the newly created container.
this fails when using pipe restore, since there is no file for the
mappings to be recovered from.
Co-Authored by: Mira Limbeck <m.limbeck@proxmox.com> Signed-off-by: Oguz Bektas <o.bektas@proxmox.com>
Oguz Bektas [Mon, 11 Feb 2019 14:51:06 +0000 (15:51 +0100)]
fix #2086: change process checking mechanism in vmstatus
vmstatus checked if the container was running by looking at the pid,
which was not an indicator of the process being completely stopped, as
the command socket in /proc/net/unix stays a little while after the
process is dead according to lxc-info.
this resulted in destroy_vm and similar functions which use
/proc/net/unix command socket based checking mechanism to fail when
executed too fast after the vm_status reported the process as stopped.
this changes vm_status to use the same kind of command socket based
mechanism in order to avoid reporting the container as being stopped too
early.
Rhonda D'Vine [Wed, 30 Jan 2019 13:41:33 +0000 (14:41 +0100)]
Fix #1924: add snapshot parameter
The pct CLI command offer the config function. The output of that may
vary with respect to a given snapshot. This adds a switch that shows the
corresponding snapshot's config.
The code needs a newer libpve-guest-common-perl, thus bumping the
dependency.
Thomas Lamprecht [Mon, 28 Jan 2019 07:06:48 +0000 (08:06 +0100)]
fix #889: api create: reserver config with create lock early
allows to remove some checks as we can be sure the config belongs to
us once we have it resered, either for restore or new creation.
This is similar to the qemu-server approach[0][1], adapted to the
LXC code. We need to cleanup a bit less if something fails, as the
LXC code path always removed the config and all created volumes in
this case, which means the 'create' reserve lock is gone too.
The early reserve on API entry, instead of doing it after forked
worker entry, allows to workaround the issues reported in #889 as
successful return from the API call means that the VMID is locked.
Thomas Lamprecht [Mon, 28 Jan 2019 07:06:47 +0000 (08:06 +0100)]
destroy_config: die if unlink fails
We use this in two places, in the cleanup path of the create/restore
API path and indirectly through PVE::LXC::destroy_lxc_container, once
again in the restore code path of the create API call, to cleanup a
CT before overwriting it with a backup if the force flag ist set. The
second time in the destroy CT API call, both times a hard error in a
erroneous cleanup is wanted.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
close #1785: whitelist namespaced lxc.sysfs.* entries
According do namespaces(7) these should be namespaced (iow.
changing these values on the host they are not propagated to
running containers), so it makes sense to whitelist them.
Note that these only work when also using
'lxc.mount.auto: proc:rw'
That should be enough for snapd on unprivileged containers.
For privileged containers we'd also need a way to not drop
the mac_admin capability - not sure we'd want that.
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
projects / pve-container.git / log