docs: backport Confluent, OpenCHAMI, and openEuler fixes from 3.x

[middelkoopt]

Backports documentation and template fixes from the 3.x branch to bring 4.x in sync for the Confluent, OpenCHAMI, and openEuler provisioner paths.

Confluent (dc35d5a)

• Replace Jinja {{ distro_id }} with shell ${distro_id} throughout Confluent templates — Jinja variables are not available at script runtime
• Detect distro_id dynamically from osdeploy importcheck output instead of assuming it from the config; this makes the import step robust across OS variants
• Improve kernel argument handling: instead of overwriting kernelargs wholesale, remove quiet and append the console/rd.shell arguments — preserves any existing args set by the OS image
• Add arch-conditional serial console (ttyAMA0 on aarch64, ttyS0 on x86_64)
• Update all four Confluent recipe YAMLs (AlmaLinux 10 and Rocky 10, both arches)

OpenCHAMI (7af8737)

• Add the OpenHPC updates repository to the compute image build configuration, so that update packages resolve correctly alongside the base OpenHPC repo
• Split the proxy placeholder comment out of the prose paragraph in proxy.md.j2 and convert it to a Jinja comment ({# #}) so it does not appear in rendered output

openEuler repository setup (778e6c3)

• Correct the description of the OpenHPC COPR ohpc_begin block in repos.md.j2: the block adds the COPR repository on the head node (required for tools such as yq used during Warewulf configuration), not just the EPOL update repo as the previous prose implied. Update the comment label accordingly.

Tested on aarch64 on qemu on a M3 for AlmaLinux

[github-actions]

🚀 CCache Statistics

Configuration	🐧 RHEL	🦊 openEuler
🖥️ x86_64 / gnu15	🔴 14.00%	🟢 71.11%
🖥️ x86_64 / intel	🔴 15.79%	-
💪 aarch64 / gnu15	🔴 14.00%	🟢 71.11%

📊 Detailed Statistics

openEuler (aarch64)

Cacheable calls:     45 /  93 (48.39%)
  Hits:              32 /  45 (71.11%)
    Direct:          32 /  32 (100.0%)
    Preprocessed:     0 /  32 ( 0.00%)
  Misses:            13 /  45 (28.89%)
Uncacheable calls:   48 /  93 (51.61%)
Local storage:
  Cache size (GiB): 0.0 / 5.0 ( 0.00%)
  Hits:              32 /  45 (71.11%)
  Misses:            13 /  45 (28.89%)

openEuler (x86_64)

Cacheable calls:     45 /  93 (48.39%)
  Hits:              32 /  45 (71.11%)
    Direct:          32 /  32 (100.0%)
    Preprocessed:     0 /  32 ( 0.00%)
  Misses:            13 /  45 (28.89%)
Uncacheable calls:   48 /  93 (51.61%)
Local storage:
  Cache size (GiB): 0.0 / 5.0 ( 0.00%)
  Hits:              32 /  45 (71.11%)
  Misses:            13 /  45 (28.89%)

RHEL (aarch64/gnu15)

Cacheable calls:    200 / 389 (51.41%)
  Hits:              28 / 200 (14.00%)
    Direct:          28 /  28 (100.0%)
    Preprocessed:     0 /  28 ( 0.00%)
  Misses:           172 / 200 (86.00%)
Uncacheable calls:  189 / 389 (48.59%)
Local storage:
  Cache size (GiB): 0.0 / 5.0 ( 0.09%)
  Hits:              28 / 200 (14.00%)
  Misses:           172 / 200 (86.00%)

RHEL (x86_64/gnu15)

Cacheable calls:    200 / 389 (51.41%)
  Hits:              28 / 200 (14.00%)
    Direct:          27 /  28 (96.43%)
    Preprocessed:     1 /  28 ( 3.57%)
  Misses:           172 / 200 (86.00%)
Uncacheable calls:  189 / 389 (48.59%)
Local storage:
  Cache size (GiB): 0.0 / 5.0 ( 0.09%)
  Hits:              28 / 200 (14.00%)
  Misses:           172 / 200 (86.00%)

RHEL (x86_64/intel)

Cacheable calls:    209 / 533 (39.21%)
  Hits:              33 / 209 (15.79%)
    Direct:          32 /  33 (96.97%)
    Preprocessed:     1 /  33 ( 3.03%)
  Misses:           176 / 209 (84.21%)
Uncacheable calls:  324 / 533 (60.79%)
Local storage:
  Cache size (GiB): 0.0 / 5.0 ( 0.09%)
  Hits:              33 / 209 (15.79%)
  Misses:           176 / 209 (84.21%)

---

🤖 Generated from workflow run 26793971052

[github-actions]

Test Results

 29 files  ±0  29 suites  ±0   46s ⏱️ -1s
 59 tests ±0  53 ✅ ±0  6 💤 ±0  0 ❌ ±0 
105 runs  ±0  97 ✅ ±0  8 💤 ±0  0 ❌ ±0 

Results for commit 778e6c3. ± Comparison against base commit e75019d.