pve-installer.git
3 days agobump version to 6.2-2 master
Thomas Lamprecht [Mon, 25 May 2020 19:12:03 +0000 (21:12 +0200)]
bump version to 6.2-2

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
3 days agopbs: create default DS
Thomas Lamprecht [Mon, 25 May 2020 19:11:58 +0000 (21:11 +0200)]
pbs: create default DS

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
4 days agowrite out interfaces address in CIDR notation
Thomas Lamprecht [Mon, 25 May 2020 19:11:39 +0000 (21:11 +0200)]
write out interfaces address in CIDR notation

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2 weeks agobump version to 6.1-5
Thomas Lamprecht [Sun, 10 May 2020 19:21:06 +0000 (21:21 +0200)]
bump version to 6.1-5

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
3 weeks agounconfigured: don't trap "bad" debug-shell exits
Thomas Lamprecht [Fri, 8 May 2020 09:21:46 +0000 (11:21 +0200)]
unconfigured: don't trap "bad" debug-shell exits

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
5 weeks agobump version to 6.1-4
Thomas Lamprecht [Fri, 24 Apr 2020 17:40:34 +0000 (19:40 +0200)]
bump version to 6.1-4

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
5 weeks agofollowup: fix file mode
Thomas Lamprecht [Fri, 24 Apr 2020 17:10:25 +0000 (19:10 +0200)]
followup: fix file mode

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
5 weeks agoimprove spice integration
Thomas Lamprecht [Fri, 24 Apr 2020 13:19:09 +0000 (15:19 +0200)]
improve spice integration

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
5 weeks agostart installer in full screen mode
Thomas Lamprecht [Fri, 24 Apr 2020 13:18:35 +0000 (15:18 +0200)]
start installer in full screen mode

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
5 weeks agounconfigured: make agetty listen on tty9
Thomas Lamprecht [Fri, 24 Apr 2020 13:17:26 +0000 (15:17 +0200)]
unconfigured: make agetty listen on tty9

it's a bit brittle and login is weird (enter root user twice, pw
doesn't matters) and once exited it's gone, but it's just for
"emergency debugging" anyway

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
5 weeks agounconfigured: allow pseudo terminals
Thomas Lamprecht [Fri, 24 Apr 2020 12:01:30 +0000 (14:01 +0200)]
unconfigured: allow pseudo terminals

allow to open a terminal in X

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2 months agobump version to 6.1-3
Thomas Lamprecht [Fri, 6 Mar 2020 09:57:57 +0000 (10:57 +0100)]
bump version to 6.1-3

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
2 months agofix behavior if zfs disks have no by-id path
Aaron Lauterer [Mon, 24 Feb 2020 08:37:23 +0000 (09:37 +0100)]
fix behavior if zfs disks have no by-id path

in some situations it is possible, that a disk does not have a
/dev/disk/by-id path. Reported cases are KVM with virtio-blk [0] and
with VMware player / workstation when using the paravirtual SCSI disk
types [1]. It can be reproduced in Proxmox VE by using virtio-blk disks
for the VM.

> The issue at hand happens because udev does not create the
> /dev/disk/by-id symlink in
> '/lib/udev/rules.d/60-persistent-storage.rules', since it does
> not have a serial-number for the device (which in the virtio-blk case
> is in turn maybe caused by qemu-server not adding one to the
> commandline).
Quoted from [2]

This regression was introduced with commit
e1b490865f750e08f6c9c6b7e162e7def9dcc411 which always replaced the disks
in the $vdev variable even if the result of the by-id lookup was
`undef`. Thus the paths did not point to any valid disk and the zpool
creation failed.

This patch replaces the disk paths with by-id paths only if they are
present.

[0] https://forum.proxmox.com/threads/installation-on-raidz-1-fails.65603/
[1] https://forum.proxmox.com/threads/not-install-promox-6-1-zfs-0-04-disk-vmware-15-pro.65269/
[2] https://pve.proxmox.com/pipermail/pve-devel/2020-February/041841.html

Signed-off-by: Aaron Lauterer <a.lauterer@proxmox.com>
3 months agoRevert "drop unused policy-disable-rc.d"
Stoiko Ivanov [Wed, 19 Feb 2020 10:54:09 +0000 (11:54 +0100)]
Revert "drop unused policy-disable-rc.d"

This reverts commit 8ee12a1c34f9b5980cb382aa2bb01855349a10a5.

`policy-rc.d` invocation is still used by deb-systemd-invoke (1p), despite
it's by now unfitting name.

The policy-rc.d script the installer placed into the chroot prevents the
starting of services by (debianized) maintainer-scripts [0].
This should reduce installation time on one hand, and on the other hand does
mask errors in maintainer-scripts, which error out if no systemd is running
(e.g. inside the chroot)

Noticed the problem, when testing an unrelated patch and running into
an aborted installtion, due to a, already fixed, glitch (missing '||true')
in pve-lxc's postinst script.

[0] https://people.debian.org/~hmh/invokerc.d-policyrc.d-specification.txt
   (did not find anything more current, contradicting the information there)

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
3 months agohtml: tree wide trailing whitespace cleanup
Thomas Lamprecht [Thu, 6 Feb 2020 11:37:44 +0000 (12:37 +0100)]
html: tree wide trailing whitespace cleanup

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
3 months agoavoid doing package specific cleanups installer
Thomas Lamprecht [Thu, 6 Feb 2020 11:35:08 +0000 (12:35 +0100)]
avoid doing package specific cleanups installer

Packages which we have under control have to ensure themself that the
correct permissions are set, else the "install on top of Debian" also
fails. It just crowd proxinstall unecessarily, so avoid to add new
ones if possible. This specific one is handled in the respective
package since quite a bit.

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
3 months agofix product names according to branding guideline
Thomas Lamprecht [Thu, 6 Feb 2020 11:32:39 +0000 (12:32 +0100)]
fix product names according to branding guideline

IOW, do not drop the whitespace between words..

[0]: https://www.proxmox.com/images/proxmox/Proxmox-Corporate-Brandguideline-2018.pdf

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
3 months agopbs: disable bridged network
Thomas Lamprecht [Thu, 6 Feb 2020 11:07:30 +0000 (12:07 +0100)]
pbs: disable bridged network

We do not host any guests or the like here, so a bridged setup is nor
really required. Disable it for now.

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
3 months agoPBS: drop obvious alpha from SVG banner
Thomas Lamprecht [Thu, 6 Feb 2020 11:04:06 +0000 (12:04 +0100)]
PBS: drop obvious alpha from SVG banner

If, we want to display the version (ISO name?) in a dynamic way for
all products.

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
3 months agoadapt installer for pbs alpha build
Oguz Bektas [Fri, 31 Jan 2020 15:11:55 +0000 (16:11 +0100)]
adapt installer for pbs alpha build

Signed-off-by: Oguz Bektas <o.bektas@proxmox.com>
3 months agodrop unused policy-disable-rc.d
Thomas Lamprecht [Sat, 1 Feb 2020 13:34:31 +0000 (14:34 +0100)]
drop unused policy-disable-rc.d

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
5 months agoumount testdir recursively in check-* and clean
Stoiko Ivanov [Fri, 29 Nov 2019 10:53:44 +0000 (11:53 +0100)]
umount testdir recursively in check-* and clean

recursively unmount testdir before removing it. This prevents the subsequent
call to try to remove files in potentially still bindmounted /proc, /sys, /dev
filesystems.

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
6 months agorestrict admin email to same format as the JSONSchema
Dominik Csapak [Thu, 28 Nov 2019 12:39:56 +0000 (13:39 +0100)]
restrict admin email to same format as the JSONSchema

Use the same regex as the 'email' format in JSONSchema (copied from there).
Otherwise, users can enter an email address which does not match our
schema and breaks the userlist API Call with a 'return value
verification failed', and one cannot rectify this via the gui
(the user has to manually edit the user.cfg or set the email via
pveum)

i noticed this while testing and using the 'insert emoji' menu in the
context menu

Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
6 months agobump version to 6.1-2
Thomas Lamprecht [Wed, 27 Nov 2019 18:20:51 +0000 (19:20 +0100)]
bump version to 6.1-2

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
6 months agofollowup: add fixme comment with buglink for dosfstools cluster size issue
Thomas Lamprecht [Wed, 27 Nov 2019 18:15:17 +0000 (19:15 +0100)]
followup: add fixme comment with buglink for dosfstools cluster size issue

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
6 months agouse by-id paths for all vdevs on pool creation
Stoiko Ivanov [Wed, 27 Nov 2019 16:06:59 +0000 (17:06 +0100)]
use by-id paths for all vdevs on pool creation

currently only the first vdev of a ZFS rpool gets created with the stable
/dev/by-id paths, all later vdevs (in a RAID0 or RAID10 setup) are left as
sdX.

By iterating over all disks in the pool and replacing their occurence with
their by-id path we get consistent and recommended names for all vdevs.

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
6 months agowipe partitiontable after early exits
Stoiko Ivanov [Wed, 27 Nov 2019 16:06:58 +0000 (17:06 +0100)]
wipe partitiontable after early exits

by wiping the partition table after the initial sanity checks regarding
minimal size and blocksize of the device, no data is destroyed for an
install that would fail in any case.

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
6 months agofix #1211: allow install on 4kn disks
Stoiko Ivanov [Wed, 27 Nov 2019 16:06:57 +0000 (17:06 +0100)]
fix #1211: allow install on 4kn disks

Installation on disks with 4k logical blocksize (4kn) failed, because
the bios_boot (a.k.a. gdisk partitiontype EF02, place for grub in
legacy BIOS boot mode) partition is created using start and end
sectors (and sector 2047 is not at 1M, where the ESP starts)

This patch addresses the issue by not creating the bios_boot
partition on 4kn disks at all - legacy boot from 4kn disks is not
supported by most BIOS implementations and grub does not support it
[0].

Checks for 4kn disks, when booted in legacy mode are added, and
prevent from leaving the harddisk selection page, if an not bootable
selection was made.

The partition numbering was kept (esp is partition 2, data is
partition 3, for consistency with other installations)

If any of the bootable disks is 4kn then the installation of the grub
legacy installation is skipped altogether.

Additionally the invocation of mkfs.vfat needs to add the parameter
'-s1' to create a bootable ESP on 4kn disks.
For a 512M vfat partition on a 512n or 512e disk, mkfs.vfat
calculates a default value of 8 sectors per cluster. For a 512M
partition on a 4kn disk, we need to scale this by 512/4096=8 and
explicitly set '-s' to 1 accordingly, since the calculation in
mkfs.vfat is brokenly assuming 512b sectors[1].

Tested with a qemu-machine by passing
'logical_block_size=4096,physical_block_size=4096' to the disk's device lines
and installing in UEFI and legacy booted mode:
* ZFS RAIDZ1
* ZFS single-disk
* ZFS RAID10 (in legacy mode grub fails to install, if any 4kn disk is in the
  pool, even if it's not in the first vdev)
* EXT4

[0] http://savannah.gnu.org/bugs/?46700
[1]: https://github.com/dosfstools/dosfstools/issues/111

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
6 months agolinewrap bootdevinfo generation
Stoiko Ivanov [Wed, 27 Nov 2019 16:06:56 +0000 (17:06 +0100)]
linewrap bootdevinfo generation

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
6 months agoadd logical_blocksize to hd_list infos
Stoiko Ivanov [Wed, 27 Nov 2019 16:06:55 +0000 (17:06 +0100)]
add logical_blocksize to hd_list infos

in preparation for 4kn installer support add the logical_blocksize to each
disk found.

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
6 months agoexpand comment on /sys/block/$dev/size
Stoiko Ivanov [Wed, 27 Nov 2019 16:06:54 +0000 (17:06 +0100)]
expand comment on /sys/block/$dev/size

/sys/block/$dev/size is also given in 512b sectors even for 4kn disks.

Since it surprised me - mention it explicitly in the comment

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
6 months agobump version to 6.1-1
Thomas Lamprecht [Wed, 20 Nov 2019 09:35:52 +0000 (10:35 +0100)]
bump version to 6.1-1

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
7 months agoensure /etc/aliases.db is generated
Stoiko Ivanov [Wed, 23 Oct 2019 16:10:27 +0000 (18:10 +0200)]
ensure /etc/aliases.db is generated

postfix relies on /etc/aliases.db for forwarding mail, without this file
all mails get deferred. (see `man newaliases`)

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
7 months agoraise postifx main.cf compatibility_level to 2
Stoiko Ivanov [Wed, 23 Oct 2019 16:10:26 +0000 (18:10 +0200)]
raise postifx main.cf compatibility_level to 2

otherwise a warning was issued (when missing /etc/aliases.db) that the system
is using the backward compatible setting of $mydestination for $relay_domain
(see [0]).

[0] http://www.postfix.org/COMPATIBILITY_README.html#relay_domains

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
9 months agobump version to 6.0-9
Thomas Lamprecht [Sat, 17 Aug 2019 08:31:51 +0000 (10:31 +0200)]
bump version to 6.0-9

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
9 months agofix #2313: load country.dat file with utf8 encoding
Dominik Csapak [Fri, 9 Aug 2019 10:47:55 +0000 (12:47 +0200)]
fix #2313: load country.dat file with utf8 encoding

this fixes the display and selection problem with countries that
have non-ascii symbols in them, like "Côte d'Ivoire"

Signed-off-by: Dominik Csapak <d.csapak@proxmox.com>
10 months agobump version to 6.0-8
Thomas Lamprecht [Thu, 25 Jul 2019 10:40:50 +0000 (12:40 +0200)]
bump version to 6.0-8

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agoefiboot: remove redundant call to zz-pve-efiboot
Fabian Grünbichler [Tue, 23 Jul 2019 12:52:47 +0000 (14:52 +0200)]
efiboot: remove redundant call to zz-pve-efiboot

this is already done by 'pve-efiboot-tool init'

Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
10 months agohandle bootloader installation errors gracefully
Fabian Grünbichler [Tue, 23 Jul 2019 12:52:46 +0000 (14:52 +0200)]
handle bootloader installation errors gracefully

by collecting them, but proceeding with the rest of the installation. it
is possible that only either legacy or EFI boot fails, and even if both
do, manually fixing the bootloader setup by booting a live CD is very
often possible.

still display the error screen at the end if bootloader setup was not
successful, so that expectations are not set too high..

Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
10 months agosuccess: show real final IP addr with port
Thomas Lamprecht [Tue, 16 Jul 2019 13:06:24 +0000 (15:06 +0200)]
success: show real final IP addr with port

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agobump version to 6.0-5
Thomas Lamprecht [Thu, 11 Jul 2019 17:22:56 +0000 (19:22 +0200)]
bump version to 6.0-5

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agouse pve-efiboot-tool for systemd boot ESP preparation
Thomas Lamprecht [Wed, 10 Jul 2019 21:35:22 +0000 (23:35 +0200)]
use pve-efiboot-tool for systemd boot ESP preparation

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agoprompt user if a vgrename is OK for exisiting 'pve'/'pmg' VGs
Thomas Lamprecht [Tue, 9 Jul 2019 19:52:37 +0000 (21:52 +0200)]
prompt user if a vgrename is OK for exisiting 'pve'/'pmg' VGs

If one has a 'pve' VG on a disks not selected as install target
(e.g., on re-installation to different disk or if putting a used disk
into another server where PVE was installed on) the vgcreate call
errored out, as for creation the VG names must be unique.

Cope with that by asking the users if a rename to a
'<vgname>-OLD-<short-uid>' name is OK in this case. It ensures that
no data is lost and that we can safely continue with the
installation. The admin can then later on wipe the renamed VG if it
was really decommissioned or save data from it (or actually use it)

This can cope (tested) with:
* a single 'pve' VG on another device
* a single 'pve' VG spanning multiple devices
* multiple 'pve' VGs spanning different sets of devices

This is achieved by using the VG UUID for rename, and by recording
all PVs with said UUID as index.

Note, while this commit message talks mostly about 'pve' VG the patch
itself is actually agnostic of the specific name, works for 'pmg' and
possible (future) other VG names too.

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agocopy /etc/hostid from installer root to target
Stoiko Ivanov [Tue, 9 Jul 2019 16:09:19 +0000 (18:09 +0200)]
copy /etc/hostid from installer root to target

/etc/hostid is used by ZFS (spl.ko) to determine which host last imported a
pool creating and importing a pool with one hostid during install and booting
with a different one (or none) leads to the system refusing to import the pool
see spl-module-parameters(5) and zpool(8).

by copying the /etc/hostid from the installer into the target system we ensure
that it stays consistent

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
10 months agofix #1603: create a new and unique machine-id
Stoiko Ivanov [Tue, 9 Jul 2019 16:09:18 +0000 (18:09 +0200)]
fix #1603: create a new and unique machine-id

see machine-id(5). The machine-id serves as a partial replacement to the hostid
(gethostid(3)) used by systemd and should be unique.

By generating a new one with `systemd-id128 new` (see machine-id(5),
sd-id128(3)) after the installation the newly installed system gets a unique
one.

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
10 months agoalso umount efivars and minor improvements
Thomas Lamprecht [Mon, 8 Jul 2019 16:03:25 +0000 (18:03 +0200)]
also umount efivars and minor improvements

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agomount efivarfs to ensure we can read bigger variables
Thomas Lamprecht [Mon, 8 Jul 2019 14:13:23 +0000 (16:13 +0200)]
mount efivarfs to ensure we can read bigger variables

In short, EFI variables can get quite big, and the old sysfs
interface was made for when they couldn't. A few firmwares out there
have such big variables, and if those are accessed through the sysfs
backed interface one gets a "Input/Output Error". 'grub-install'
chokes on that error when it iterates over all variables to do it's
stuff, and thus fails our installation. When we mount the efivarfs,
which does not has this limitations, one can read all variables just
fine - at least as long as the NVRAM backing them is not broken.

from Linux Kernel Documentation/filesystems/efivarfs.txt:
> The efivarfs filesystem was created to address the shortcomings of
> using entries in sysfs to maintain EFI variables. The old sysfs EFI
> variables code only supported variables of up to 1024 bytes. This
> limitation existed in version 0.99 of the EFI specification, but was
> removed before any full releases. Since variables can now be larger
> than a single page, sysfs isn't the best interface for this.
> Variables can be created, deleted and modified with the efivarfs
> filesystem.

Also mount it in the installer environment for debugging purpose.

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agounderline that we detected less than 1024 MB of *useable* memory
Thomas Lamprecht [Fri, 5 Jul 2019 13:43:02 +0000 (15:43 +0200)]
underline that we detected less than 1024 MB of *useable* memory

firmwar, kernel, and some PCI(e) devices can use some memory, which
then is not available for userspace, thus the total memory installed
for this server normally differs at least about 50 MB from the
useable one.

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agohd_list: skip all devices with iso9660 as FS
Thomas Lamprecht [Fri, 5 Jul 2019 13:21:10 +0000 (15:21 +0200)]
hd_list: skip all devices with iso9660 as FS

This ensures that when booting from a USB stick, or similar device,
we do not show that device in the "Target Harddisk" selector.

While there may be some constructed cases where this actually worked
and was wanted, e.g., one created a partition on the end of a USB
storage device, put the ISO on that and used the beginning of that
device as target, I'd say that even if that worked with luck, it's
better not actively not support it.

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agobump version to 6.0-4
Thomas Lamprecht [Thu, 4 Jul 2019 16:10:36 +0000 (18:10 +0200)]
bump version to 6.0-4

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agosummary: s/drive/drive(s)/
Thomas Lamprecht [Thu, 4 Jul 2019 16:12:06 +0000 (18:12 +0200)]
summary: s/drive/drive(s)/

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agoadd $product_fullname hash and use it
Thomas Lamprecht [Wed, 3 Jul 2019 08:21:49 +0000 (10:21 +0200)]
add $product_fullname hash and use it

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agofollowup: refword, coding style and do not die
Thomas Lamprecht [Wed, 3 Jul 2019 05:32:16 +0000 (07:32 +0200)]
followup: refword, coding style and do not die

If a user really want's to try it, why not.
Also, no point in dividing by 1024 if we have a less-than operator

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agoabort installation if memory is less than 1GB
Oguz Bektas [Tue, 2 Jul 2019 13:17:32 +0000 (15:17 +0200)]
abort installation if memory is less than 1GB

show a message on screen about memory requirement, then die and abort
the installation. (Note: followup removes the die again - Thomas)

The "magical" limit seems around 850 - 900 MiB memory needed, it's
a bit strange, as we do not need that much, proxinstall + webkit +
base system are around 300 MiB, the rest is page cache, but it seems
that the memory pressure gets to big.

Signed-off-by: Oguz Bektas <o.bektas@proxmox.com>
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agofix typo: s/virtualisation/virtualization/
Thomas Lamprecht [Wed, 3 Jul 2019 05:31:31 +0000 (07:31 +0200)]
fix typo: s/virtualisation/virtualization/

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agouse by-id diskpath for zpool create
Stoiko Ivanov [Tue, 2 Jul 2019 14:44:32 +0000 (16:44 +0200)]
use by-id diskpath for zpool create

With the recent fixes to the installers initial /dev creation, the installer
and the kernel should now have the same view on the created devices and links

10 months agoclear zpool labels on selected disks' partitions
Stoiko Ivanov [Tue, 2 Jul 2019 14:44:31 +0000 (16:44 +0200)]
clear zpool labels on selected disks' partitions

along with `pvremove -ff` and writing zeroes to the first 16M of all
partitions, which belong to disks selected in the installer run `zpool
labelclear`.

This prevents a failure to boot after installation, if the disks were
previously also used as a zpool called rpool, but in a different configuration
(e.g. installing a raidzX first and then changing to raid10).

10 months agofollowup: reword summary sentence a bit
Thomas Lamprecht [Tue, 2 Jul 2019 12:15:23 +0000 (14:15 +0200)]
followup: reword summary sentence a bit

change the really short sentence to something with more "reading
flow"

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agoalign lines in summary screen
Oguz Bektas [Tue, 2 Jul 2019 10:03:26 +0000 (12:03 +0200)]
align lines in summary screen

Signed-off-by: Oguz Bektas <o.bektas@proxmox.com>
10 months agofollowup: only do zz-pve-efiboot if ZFS and EFI is used, for now
Thomas Lamprecht [Mon, 1 Jul 2019 20:36:59 +0000 (22:36 +0200)]
followup: only do zz-pve-efiboot if ZFS and EFI is used, for now

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
10 months agoupdate systemd-boot config after initializing esps
Stoiko Ivanov [Mon, 1 Jul 2019 20:16:22 +0000 (22:16 +0200)]
update systemd-boot config after initializing esps

run the kernel-postinst hook once (like we do for update-grub) instead
of once per ESP. This fixes an error (the hook does not write entries for
mounted partitions), and save quite a bit of (un)mounting and copying in case
of multiple ESPs

Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
11 months agobump version to 6.0-3
Thomas Lamprecht [Fri, 28 Jun 2019 12:03:28 +0000 (14:03 +0200)]
bump version to 6.0-3

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
11 months agorefactor out prepare_grub_efi_boot_esp code
Thomas Lamprecht [Fri, 28 Jun 2019 11:02:27 +0000 (13:02 +0200)]
refactor out prepare_grub_efi_boot_esp code

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
11 months agominor coding style cleanups
Thomas Lamprecht [Fri, 28 Jun 2019 11:02:08 +0000 (13:02 +0200)]
minor coding style cleanups

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
11 months agodepreacate grub_platform variable, replaced by $boot_type
Thomas Lamprecht [Fri, 28 Jun 2019 11:01:26 +0000 (13:01 +0200)]
depreacate grub_platform variable, replaced by $boot_type

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
11 months agoadd support for ZFS on EFI
Stoiko Ivanov [Thu, 27 Jun 2019 18:27:59 +0000 (20:27 +0200)]
add support for ZFS on EFI

use systemd-boot in case of ZFS on EFI and prepare the ESPs of all bootable
devices by installing systemd-boot to them

11 months agobump version to 6.0-2
Thomas Lamprecht [Tue, 25 Jun 2019 08:28:25 +0000 (10:28 +0200)]
bump version to 6.0-2

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
11 months agoHD options panel: add seprator before OK button
Thomas Lamprecht [Wed, 19 Jun 2019 08:09:17 +0000 (10:09 +0200)]
HD options panel: add seprator before OK button

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
11 months agoZFS: add note regarding HW raid controllers
Thomas Lamprecht [Wed, 19 Jun 2019 08:07:46 +0000 (10:07 +0200)]
ZFS: add note regarding HW raid controllers

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
11 months agocreate_ipconf_view: if undefined variable in string cmp warning
Thomas Lamprecht [Fri, 7 Jun 2019 18:53:52 +0000 (20:53 +0200)]
create_ipconf_view: if undefined variable in string cmp warning

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
12 months agobuildsys: switch upload dist over to buster
Thomas Lamprecht [Wed, 22 May 2019 08:21:49 +0000 (10:21 +0200)]
buildsys: switch upload dist over to buster

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
12 months agobump version to 6.0-1
Thomas Lamprecht [Wed, 22 May 2019 08:03:21 +0000 (10:03 +0200)]
bump version to 6.0-1

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
12 months agoupdate banners
Thomas Lamprecht [Wed, 22 May 2019 08:02:55 +0000 (10:02 +0200)]
update banners

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
12 months agomount /run as tmpfs in chroot and bindmount to target root
Thomas Lamprecht [Wed, 22 May 2019 08:11:55 +0000 (10:11 +0200)]
mount /run as tmpfs in chroot and bindmount to target root

ensures things like LVM can work fine

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
12 months agoproxinstall: remove outdated cleanup code
Thomas Lamprecht [Wed, 22 May 2019 07:30:51 +0000 (09:30 +0200)]
proxinstall: remove outdated cleanup code

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
12 months agodrop also into debug shell before even starting to install
Thomas Lamprecht [Wed, 22 May 2019 07:27:52 +0000 (09:27 +0200)]
drop also into debug shell before even starting to install

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
12 months agobuildsys: use dpkg-dev makefile helpers for pkg info
Thomas Lamprecht [Wed, 22 May 2019 08:00:20 +0000 (10:00 +0200)]
buildsys: use dpkg-dev makefile helpers for pkg info

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
12 months agoignore non-existent eula file in test mode
Thomas Lamprecht [Wed, 22 May 2019 07:49:38 +0000 (09:49 +0200)]
ignore non-existent eula file in test mode

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
12 months agobe compatible with new Gtk3 Webkit2
Thomas Lamprecht [Wed, 22 May 2019 07:24:05 +0000 (09:24 +0200)]
be compatible with new Gtk3 Webkit2

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
14 months agod/control: pass ${perl:Depends} to Depends field
Thomas Lamprecht [Tue, 26 Mar 2019 14:38:29 +0000 (15:38 +0100)]
d/control: pass ${perl:Depends} to Depends field

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
14 months agounconfigured: mount /dev/shm as tmpfs inside chroot
Thomas Lamprecht [Mon, 25 Mar 2019 20:18:23 +0000 (21:18 +0100)]
unconfigured: mount /dev/shm as tmpfs inside chroot

as preparation for GKT3 WebKit2 perl bindings, which wants to make a
shared memory file there, but fails and hangs the installer as
/dev/shm didn't exists in the chroot environment.

Doesn't hurt to just create it now already

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
14 months agotell udev to ignore being in a chroot
Thomas Lamprecht [Mon, 25 Mar 2019 18:47:05 +0000 (19:47 +0100)]
tell udev to ignore being in a chroot

since udev/systemd v241 udev actively ignores requests if it detects
running in a chroot environment[0], as we do exactly this in our
setup process we need to tell udev that it is OK and we know what we
do, so set SYSTEMD_IGNORE_CHROOT=1 which makes the
"running_in_chroot" function lie[1].

[0]: https://github.com/systemd/systemd/commit/c494b739a47359ab2697482f52545e2a6d1c86ad
[1]: https://github.com/systemd/systemd/blob/v241/src/basic/virt.c#L607

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
14 months agobump version 5.0-26
Thomas Lamprecht [Wed, 20 Mar 2019 13:43:40 +0000 (14:43 +0100)]
bump version 5.0-26

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
14 months agoipconf view: only (re)set fields if nic really changed
Thomas Lamprecht [Wed, 20 Mar 2019 13:38:04 +0000 (14:38 +0100)]
ipconf view: only (re)set fields if nic really changed

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
14 months agobump version to 5.0-25
Thomas Lamprecht [Thu, 14 Mar 2019 15:36:52 +0000 (16:36 +0100)]
bump version to 5.0-25

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
14 months agofixup comment indentation
Thomas Lamprecht [Thu, 7 Mar 2019 06:16:41 +0000 (07:16 +0100)]
fixup comment indentation

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
14 months agoAdd support for Citrix Hypervisor blockdev.
Wolfgang Link [Wed, 6 Mar 2019 14:52:58 +0000 (15:52 +0100)]
Add support for Citrix Hypervisor blockdev.

Citrix Hypervisor use as blockdev path /dev/xvd<[a-z]><\d>
It makes sense for Proxmox Mail Gateway to run on Citrix Hypervisor.

Citrix Hypervisor recent XenServer.

16 months agos/$i/$a/
Thomas Lamprecht [Fri, 25 Jan 2019 15:28:27 +0000 (16:28 +0100)]
s/$i/$a/

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
16 months agofixup-follow-up
Thomas Lamprecht [Fri, 25 Jan 2019 15:26:18 +0000 (16:26 +0100)]
fixup-follow-up

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
16 months agofollowup: really remember last HD selected
Thomas Lamprecht [Fri, 25 Jan 2019 15:24:15 +0000 (16:24 +0100)]
followup: really remember last HD selected

This was done correctly earlier, but the remembering part disappeared
on rebase, so followup, but do not save it in config_options.

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
16 months agodon't show hd select combobox after raid is chosen
Oguz Bektas [Fri, 25 Jan 2019 15:14:02 +0000 (16:14 +0100)]
don't show hd select combobox after raid is chosen

* we don't want the hd select combobox to show up when user goes back
after choosing raid during initial setup

Signed-off-by: Oguz Bektas <o.bektas@proxmox.com>
16 months agoonly set target_hd if it's not defined already
Oguz Bektas [Fri, 25 Jan 2019 15:14:01 +0000 (16:14 +0100)]
only set target_hd if it's not defined already

* this avoids weird behaviour with the chosen disk when the user changes
the type of filesystem during installation

Signed-off-by: Oguz Bektas <o.bektas@proxmox.com>
16 months agofix typo in variable name
Oguz Bektas [Fri, 25 Jan 2019 15:14:00 +0000 (16:14 +0100)]
fix typo in variable name

Signed-off-by: Oguz Bektas <o.bektas@proxmox.com>
16 months agofixup: remove unused wrongly introduced variable
Thomas Lamprecht [Fri, 11 Jan 2019 17:54:12 +0000 (18:54 +0100)]
fixup: remove unused wrongly introduced variable

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
16 months agoack template: s/Disk/Disk(s)/
Thomas Lamprecht [Fri, 11 Jan 2019 13:05:13 +0000 (14:05 +0100)]
ack template: s/Disk/Disk(s)/

As we show all selected disk in a ZFS/BTRFS raid mode

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
16 months agocreate_hdsel_view: cleanup set_next
Thomas Lamprecht [Fri, 11 Jan 2019 12:59:38 +0000 (13:59 +0100)]
create_hdsel_view: cleanup set_next

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
16 months agoshow multiple disks in ack screen when zfs is used
Thomas Lamprecht [Thu, 10 Jan 2019 17:09:04 +0000 (18:09 +0100)]
show multiple disks in ack screen when zfs is used

Tested-by: Oguz Bektas <o.bektas@proxmox.com>
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
17 months agodo not create resolvconf links
Thomas Lamprecht [Fri, 14 Dec 2018 15:13:12 +0000 (16:13 +0100)]
do not create resolvconf links

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
17 months agoack screen: show real country name instead of 2 letter code
Thomas Lamprecht [Fri, 14 Dec 2018 10:25:37 +0000 (11:25 +0100)]
ack screen: show real country name instead of 2 letter code

Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com>
17 months agominor style fixes
Oguz Bektas [Fri, 14 Dec 2018 09:42:56 +0000 (10:42 +0100)]
minor style fixes

Signed-off-by: Oguz Bektas <o.bektas@proxmox.com>