[pve-devel] [PATCH container] fix #4460: setup: centos: create /etc/hostname if it does not exist

[pve-devel] [PATCH container] fix #4460: setup: centos: create /etc/hostname if it does not exist

[Friedrich Weber]

Previously, Setup/CentOS.pm only wrote to /etc/hostname if the file
already existed. Many CT templates of Redhat-derived distros do not
contain that file, so the containers ended up without /etc/hostname.
This caused systemd-hostnamed to report the "static hostname" to be
empty. If networking is handled by NetworkManager, the empty static
hostname caused DHCP requests to be sent without the "Hostname" field,
as reported in #4460.

With this fix, Setup/CentOS.pm creates /etc/hostname if it does not
exist, so NetworkManager correctly reads the hostname and includes it in
DHCP requests.

Manually tested with the following CT templates (checking that
/etc/hostname exists and DHCP requests include the hostname):

* Distros using NetworkManager:
 - Alma Linux 9 (almalinux-9-default_20221108_amd64.tar.xz)
 - CentOS 8 (centos-8-default_20201210_amd64.tar.xz)
 - CentOS 9 Stream (centos-9-stream-default_20221109_amd64.tar.xz)
 - Rocky Linux 9 (rockylinux-9-default_20221109_amd64.tar.xz)
* Distros using network-scripts (here, DHCP requests already contained the
hostname without this fix, as network-scripts does not rely on
systemd-hostnamed):
 - Alma Linux 8 (almalinux-8-default_20210928_amd64.tar.xz)
 - CentOS 7 (centos-7-default_20190926_amd64.tar.xz)
 - CentOS 8 Stream (centos-8-stream-default_20220327_amd64.tar.xz)
 - Rocky Linux 8 (rockylinux-8-default_20210929_amd64.tar.xz)

Signed-off-by: Friedrich Weber <f.weber@proxmox.com>
---

Question: This will cause Setup/CentOS.pm to create /etc/hostname also
in already-existing containers. I don't think this should any cause
issues for users, but I'm not sure. What do you think?

 src/PVE/LXC/Setup/CentOS.pm                | 5 ++---
 src/test/test-centos6-001/etc/hostname.exp | 1 +
 src/test/test-centos6-002/etc/hostname.exp | 1 +
 src/test/test-centos8-001/etc/hostname.exp | 1 +
 4 files changed, 5 insertions(+), 3 deletions(-)
 create mode 100644 src/test/test-centos6-001/etc/hostname.exp
 create mode 100644 src/test/test-centos6-002/etc/hostname.exp
 create mode 100644 src/test/test-centos8-001/etc/hostname.exp

diff --git a/src/PVE/LXC/Setup/CentOS.pm b/src/PVE/LXC/Setup/CentOS.pm
index 00fecc6..1d31cee 100644
--- a/src/PVE/LXC/Setup/CentOS.pm
+++ b/src/PVE/LXC/Setup/CentOS.pm
@@ -157,9 +157,8 @@ sub set_hostname {
 
     $self->update_etc_hosts($hostip, $oldname, $hostname, $searchdomains);
 
-    if ($self->ct_file_exists($hostname_fn)) {
-	$self->ct_file_set_contents($hostname_fn, "$hostname\n");
-    }
+    # Always write /etc/hostname, even if it does not exist yet
+    $self->ct_file_set_contents($hostname_fn, "$hostname\n");
 
     if ($self->ct_file_exists($sysconfig_network)) {
 	my $data = $self->ct_file_get_contents($sysconfig_network);
diff --git a/src/test/test-centos6-001/etc/hostname.exp b/src/test/test-centos6-001/etc/hostname.exp
new file mode 100644
index 0000000..a5bce3f
--- /dev/null
+++ b/src/test/test-centos6-001/etc/hostname.exp
@@ -0,0 +1 @@
+test1
diff --git a/src/test/test-centos6-002/etc/hostname.exp b/src/test/test-centos6-002/etc/hostname.exp
new file mode 100644
index 0000000..180cf83
--- /dev/null
+++ b/src/test/test-centos6-002/etc/hostname.exp
@@ -0,0 +1 @@
+test2
diff --git a/src/test/test-centos8-001/etc/hostname.exp b/src/test/test-centos8-001/etc/hostname.exp
new file mode 100644
index 0000000..a5bce3f
--- /dev/null
+++ b/src/test/test-centos8-001/etc/hostname.exp
@@ -0,0 +1 @@
+test1
-- 
2.30.2

[pve-devel] applied: [PATCH container] fix #4460: setup: centos: create /etc/hostname if it does not exist

[Thomas Lamprecht]

Am 16/01/2023 um 17:52 schrieb Friedrich Weber:
> Previously, Setup/CentOS.pm only wrote to /etc/hostname if the file

Nit: note that we prefer avoiding code file names in favor of module names
or more general terms like "the CentOS setup module" - if one wants to
know which file got altered they can always check the git diff/show --stat
output. There are naturally some commits where it may be warranted, but
it's rather an exception.

The rest of the commit message and patch was really great, so it's really
just an very minor nit.

> already existed. Many CT templates of Redhat-derived distros do not
> contain that file, so the containers ended up without /etc/hostname.
> This caused systemd-hostnamed to report the "static hostname" to be
> empty. If networking is handled by NetworkManager, the empty static
> hostname caused DHCP requests to be sent without the "Hostname" field,
> as reported in #4460.
> 
> With this fix, Setup/CentOS.pm creates /etc/hostname if it does not
> exist, so NetworkManager correctly reads the hostname and includes it in
> DHCP requests.
> 
> Manually tested with the following CT templates (checking that
> /etc/hostname exists and DHCP requests include the hostname):
> 
> * Distros using NetworkManager:
>  - Alma Linux 9 (almalinux-9-default_20221108_amd64.tar.xz)
>  - CentOS 8 (centos-8-default_20201210_amd64.tar.xz)
>  - CentOS 9 Stream (centos-9-stream-default_20221109_amd64.tar.xz)
>  - Rocky Linux 9 (rockylinux-9-default_20221109_amd64.tar.xz)
> * Distros using network-scripts (here, DHCP requests already contained the
> hostname without this fix, as network-scripts does not rely on
> systemd-hostnamed):
>  - Alma Linux 8 (almalinux-8-default_20210928_amd64.tar.xz)
>  - CentOS 7 (centos-7-default_20190926_amd64.tar.xz)
>  - CentOS 8 Stream (centos-8-stream-default_20220327_amd64.tar.xz)
>  - Rocky Linux 8 (rockylinux-8-default_20210929_amd64.tar.xz)
> 
> Signed-off-by: Friedrich Weber <f.weber@proxmox.com>
> ---
> 
> Question: This will cause Setup/CentOS.pm to create /etc/hostname also
> in already-existing containers. I don't think this should any cause
> issues for users, but I'm not sure. What do you think?

first, as long as the central ct_file_set_contents is used any admin/user
can control if a file is ignored by creating a ".pve-ignore.$file" in the
same directory.

When initially introducing this in c0eae40 ("add class for redhat based
containers") we can see that the if was accompanied by an else earlier, as
it then was deemed to either set the hostname via /etc/hostname or the
sysconfig. There's no actual reason for why that's the case, as after
all it would always be the same value, so it seems that it couldn't
have ever hurt to set both, well, at least if nothing changed behavior
significantly depending on the existence of /etc/hostname - which would
be IMO really weird and in any case quite hard to find out for all possible
CentOS et al. packages - so rather I'd just roll this out slowly and check
for regression reports in the support channels.

> 
>  src/PVE/LXC/Setup/CentOS.pm                | 5 ++---
>  src/test/test-centos6-001/etc/hostname.exp | 1 +
>  src/test/test-centos6-002/etc/hostname.exp | 1 +
>  src/test/test-centos8-001/etc/hostname.exp | 1 +
>  4 files changed, 5 insertions(+), 3 deletions(-)
>  create mode 100644 src/test/test-centos6-001/etc/hostname.exp
>  create mode 100644 src/test/test-centos6-002/etc/hostname.exp
>  create mode 100644 src/test/test-centos8-001/etc/hostname.exp
> 
>

applied, thanks!