Files
PSProxmoxVE/DECISIONS.md
T
goodolclint-claude[bot] fc073e7e2a ci: pin Pester by exact version everywhere it is installed or imported
Pester was installed with -MinimumVersion 5.0 and no ceiling in the CI job
image, both install steps in unit-tests.yml, and both Import-Module calls, plus
the suite's own import inside the container. The image is rebuilt on every CI
run and Pester is installed fresh on every unit-test run, so PSGallery chose the
version — a new major could reach the required PR checks with no commit here,
surfacing as unexplained test breakage on whichever PR ran next.

It had already happened. Steps named "Install Pester 5" were resolving 6.1.0 on
both legs, because Pester 6 declares PowerShellVersion 5.1 and so installs on
Windows PowerShell too. Nothing broke — the suite uses only constructs common to
5 and 6, and runs 1566/0 under 6.1.0 with no deprecation warnings — but nobody
chose it. The step names are corrected; they had been describing an install that
stopped happening some time ago.

Pinning the install alone is not enough, in two ways review found:

An unset variable does not fail. -RequiredVersion accepts an empty value and
degrades to "latest" for Install-Module and to "any" for Import-Module, both
exiting 0, so a renamed or dropped env key would silently restore the float this
commit removes. A guard step now fails the job instead.

The point of use was still floored. run-integration.sh imported the suite's
Pester with -MinimumVersion 5.0, so a second Pester reaching PSModulePath would
win regardless of what was installed. The Dockerfile now promotes the ARG to ENV
so the version is discoverable at runtime, and that import is pinned to it.

The pin lives in two files, so shell-selfchecks asserts they agree — split-brain
between the workflow and the image is precisely the unexplained breakage this is
meant to prevent. CONTRIBUTING.md and CLAUDE.md are updated too; the contributor
instructions were a third floating install site.

Recorded as an amendment to D017 — the same principle as the nested PVE package
pin, applied to the lane's own tooling.
2026-09-01 18:08:02 -05:00

30 KiB
Raw Blame History

Architectural Decisions

This file documents patterns that were deliberately chosen or changed, anti-patterns that were explicitly removed, and constraints that must be maintained. Read this before writing any new code.


D001 — Task polling must use TaskService.WaitForTask

Status: Active Finding refs: F032, F033, F036, F058 Resolved in scan: 2026-03-22 (for VM/network cmdlets); container snapshot + storage cmdlets still open

Decision

All task-polling loops must use TaskService.WaitForTask(upid, session, timeout, progress). Never implement inline while(true) or do/while polling loops in cmdlet files.

Rationale

Four VM/network cmdlets (InvokePveNetworkApply, NewPveSnapshot, RestorePveSnapshot, RemovePveSnapshot) and one guest exec cmdlet had copy-pasted polling loops with no timeout, causing cmdlets to hang indefinitely if a PVE task stalled. TaskService.WaitForTask has timeout enforcement, failure detection, and WriteProgress support.

Five additional cmdlets (3 container snapshot + 2 storage) still have this anti-pattern as of scan 2026-03-22 (F058).

Anti-pattern (do not reintroduce)

// NEVER do this in a cmdlet
while (true)
{
    var status = taskService.GetTask(upid, session);
    if (status.IsFinished) break;
    Thread.Sleep(1000);
}

Correct pattern

// Always use this
TaskService.WaitForTask(upid, session, TimeoutSeconds, this);

D002 — Password parameters must use SecureString

Status: Active Finding refs: F051 Resolved in scan: 2026-03-22

Decision

All cmdlet parameters that accept passwords must use SecureString type with Marshal.SecureStringToGlobalAllocUnicode + ZeroFreeGlobalAllocUnicode in a try/finally block for extraction.

Rationale

Set-PveVmGuestPassword originally accepted a plain string password parameter, leaving the credential in managed memory indefinitely. SecureString minimizes the window of exposure and is consistent with Connect-PveServer's PSCredential handling.

Anti-pattern (do not reintroduce)

// NEVER accept passwords as plain strings
[Parameter(Mandatory = true)]
public string Password { get; set; }

Correct pattern

[Parameter(Mandatory = true)]
public SecureString Password { get; set; }

// In ProcessRecord:
IntPtr ptr = IntPtr.Zero;
try
{
    ptr = Marshal.SecureStringToGlobalAllocUnicode(Password);
    string plainText = Marshal.PtrToStringUni(ptr);
    // Use plainText for API call
}
finally
{
    if (ptr != IntPtr.Zero)
        Marshal.ZeroFreeGlobalAllocUnicode(ptr);
}

D003 — URL encoding required for all path parameters

Status: Active Finding refs: F050 Resolved in scan: 2026-03-22

Decision

All user-supplied or dynamic values interpolated into API URL paths must be wrapped in Uri.EscapeDataString(). This applies to all service classes.

Rationale

Snapshot names, node names, user IDs, and other identifiers could theoretically contain characters that break URL path segments. While most values come from validated sources, defense-in-depth requires consistent encoding. Applied across all 14 service classes.

Anti-pattern (do not reintroduce)

// NEVER interpolate raw strings into URL paths
var resource = $"nodes/{node}/qemu/{vmid}/snapshot/{snapshotName}";

Correct pattern

var resource = $"nodes/{Uri.EscapeDataString(node)}/qemu/{vmid}/snapshot/{Uri.EscapeDataString(snapshotName)}";

D004 — No bare catch blocks

Status: Active Finding refs: F039 Resolved in scan: 2026-03-22

Decision

No bare catch { } or catch (Exception) { } blocks. All catch blocks must either:

  1. Use a specific exception type (catch (PveApiException ex)), or
  2. Use a filtered catch with when clause that excludes fatal exceptions (catch (Exception ex) when (ex is not OutOfMemoryException and not StackOverflowException))

Rationale

Bare catches in PveHttpClient, PveCmdletBase, VmService, ContainerService, and GetPveVmCmdlet silently swallowed errors, making debugging impossible. Replacing with filtered or specific catches preserves error visibility while still handling expected transient failures.

Anti-pattern (do not reintroduce)

// NEVER use bare catches
try { ... }
catch { }

// NEVER catch all exceptions unfiltered
try { ... }
catch (Exception) { /* ignore */ }

Correct pattern

// Catch specific exceptions
try { ... }
catch (PveApiException ex) { WriteWarning(ex.Message); }

// Or use filtered catch for status polling
try { ... }
catch (Exception ex) when (ex is not OutOfMemoryException and not StackOverflowException)
{
    WriteVerbose($"Status poll failed: {ex.Message}");
}

D005 — OutputType required on all cmdlets

Status: Active Finding refs: F037 Resolved in scan: 2026-03-22

Decision

Every cmdlet must have an [OutputType(typeof(...))] attribute declaring its return type.

Rationale

~54 cmdlets were missing OutputType, degrading IntelliSense, pipeline type inference, and Get-Command -OutputType queries. All 169 cmdlets now have the attribute.

Correct pattern

[Cmdlet(VerbsCommon.Get, "PveVm")]
[OutputType(typeof(VmInfo))]
public sealed class GetPveVmCmdlet : PveCmdletBase

D006 — ConfirmImpact.High required for destructive operations

Status: Active Finding refs: F011, F034, F042, F043, F062, F063 Resolved in scan: 2026-03-22 (for VM cmdlets); container Restart/Suspend still open

Decision

All cmdlets that perform destructive or disruptive operations must set ConfirmImpact = ConfirmImpact.High in the [Cmdlet] attribute. This includes:

  • All Remove-* cmdlets
  • All Stop-* cmdlets
  • All Reset-* cmdlets
  • All Restart-* cmdlets
  • All Suspend-* cmdlets
  • Restore-PveSnapshot and Restore-PveContainerSnapshot
  • New-PveTemplate (irreversible conversion)

Rationale

Stop-PveVm, Reset-PveVm, Suspend-PveVm, Restart-PveVm, and Remove-PveRole were missing ConfirmImpact.High, meaning users could accidentally perform disruptive operations without being prompted. Container counterparts (Restart/Suspend) remain inconsistent as of F062/F063.

Correct pattern

[Cmdlet(VerbsLifecycle.Stop, "PveVm", SupportsShouldProcess = true,
    ConfirmImpact = ConfirmImpact.High)]

D007 — All cmdlet classes must be sealed

Status: Active Finding refs: F041 Resolved in scan: 2026-03-22

Decision

All cmdlet classes must be declared sealed. Cmdlets are not designed for inheritance and sealing prevents unintended extension.

Rationale

~95 cmdlets were not sealed. Sealing all 169 cmdlets makes the design intent explicit and enables potential JIT optimizations.

Correct pattern

public sealed class GetPveVmCmdlet : PveCmdletBase

D008 — JSON serialization: Newtonsoft.Json only

Status: Active Finding refs: F044 Resolved in scan: 2026-03-22

Decision

Use only Newtonsoft.Json ([JsonProperty]) for JSON serialization attributes on model classes. Do not add System.Text.Json ([JsonPropertyName]) attributes.

Rationale

The module uses Newtonsoft.Json for all API response deserialization. Having both [JsonProperty] and [JsonPropertyName] attributes was redundant and confusing — System.Text.Json is not used at runtime. All [JsonPropertyName] attributes were removed.

Anti-pattern (do not reintroduce)

// NEVER add System.Text.Json attributes alongside Newtonsoft
[JsonProperty("status")]
[JsonPropertyName("status")]  // Don't add this
public string Status { get; set; }

Correct pattern

[JsonProperty("status")]
public string Status { get; set; }

D009 — Framework targeting: netstandard2.0 for publishable, net10.0+net48 for tests

Status: Active Finding refs: F047, F064 Status note: net9.0 → net10.0 migration still pending

Decision

  • Publishable projects (PSProxmoxVE, PSProxmoxVE.Core): Target netstandard2.0 for maximum compatibility (PS 5.1 Desktop + PS 7.x Core).
  • Test projects: Target net10.0 (LTS) and net48 (Windows PowerShell 5.1 validation).

Rationale

.NET 9.0 reached EOL in May 2025. The test projects should use the current LTS release (net10.0). The publishable module must remain on netstandard2.0 to support both Desktop and Core editions.


D010 — VmId parameters: nullable int with ValidateRange

Status: Active Finding refs: F012, F038 Resolved in scan: 2026-03-21 (ValidateRange), 2026-03-22 (nullable)

Decision

VmId parameters must:

  1. Use int? (nullable) when the parameter is optional (e.g., firewall cmdlets that operate at cluster, node, or VM level)
  2. Include [ValidateRange(100, 999999999)] to match PVE's VMID constraints
  3. Use int (non-nullable) only when VmId is mandatory

Rationale

PVE requires VMIDs in range 100-999999999. Without ValidateRange, invalid IDs reach the API and return confusing errors. Using non-nullable int with default 0 for optional VmId made it impossible to distinguish "not specified" from "VM 0" in firewall cmdlets.

Correct pattern

// Mandatory VmId
[Parameter(Mandatory = true)]
[ValidateRange(100, 999999999)]
public int VmId { get; set; }

// Optional VmId (e.g., firewall cmdlets)
[Parameter()]
[ValidateRange(100, 999999999)]
public int? VmId { get; set; }

D011 — Verb class constants required for cmdlet attributes

Status: Active Finding refs: F009 Resolved in scan: 2026-03-21

Decision

All [Cmdlet] attributes must use verb class constants (VerbsCommon.Get, VerbsLifecycle.Start, etc.) instead of hardcoded string literals.

Rationale

Reset-PveVm used [Cmdlet("Reset", ...)] instead of VerbsCommon.Reset. While "Reset" is an approved verb, using the constant ensures compile-time verification and consistency with all other cmdlets.

Anti-pattern (do not reintroduce)

[Cmdlet("Reset", "PveVm")]  // Don't use string literals

Correct pattern

[Cmdlet(VerbsCommon.Reset, "PveVm")]

D012 — Magic strings: extract to named constants

Status: Active Finding refs: F049 Resolved in scan: 2026-03-22

Decision

Frequently used string literals (auth header names, token prefixes, etc.) should be extracted to const string fields for maintainability.

Rationale

Auth header names (PVEAPIToken=, CSRFPreventionToken) were inline string literals used in multiple places. Extracting to const string ApiTokenPrefix and CsrfHeaderName fields improves maintainability and reduces typo risk.


D013 — Cmdlets must emit only native or module-defined types

Status: Active Finding refs: F085 Resolved in scan: 2026-03-25

Decision

All cmdlet output types and public model properties must be native .NET types (string, int, bool, Dictionary<string, object?>, List<T>, PSObject, void) or types defined within the module itself (Pve* classes). Never expose third-party types like Newtonsoft's JObject, JArray, or JToken in public APIs.

Rationale

PowerShell enumerates JArray unexpectedly and JObject properties are not discoverable via Get-Member or tab completion. Users piping module output into Format-Table, Select-Object, or Where-Object get confusing behavior when the underlying type is a Newtonsoft container. Native dictionaries and lists work naturally in PowerShell pipelines.

Anti-pattern (do not reintroduce)

// Service returning Newtonsoft type
public JObject GetNodeConfig(...) { ... }

// Model exposing Newtonsoft type
[JsonProperty("members")]
public JArray? Members { get; set; }

// Cmdlet OutputType referencing Newtonsoft type
[OutputType(typeof(JObject))]

Correct pattern

// Service returns native dictionary
public Dictionary<string, object?> GetNodeConfig(...) { ... }

// Model uses native type with converter for deserialization
[JsonProperty("members")]
[JsonConverter(typeof(NativeListConverter))]
public List<Dictionary<string, object?>>? Members { get; set; }

// Cmdlet OutputType uses native or module type
[OutputType(typeof(Dictionary<string, object>))]
// or
[OutputType(typeof(PSObject))]

D014 — New-PveCluster -Wait blocks until the cluster is quorate

Status: Active Finding refs: (none — found via integration run 172, 2026-09-01) Resolved in scan: n/a

Decision

New-PveCluster -Wait returns only after the cluster reports quorum, not merely when the create task completes. ClusterConfigService.WaitForQuorum implements the wait; it polls GET /cluster/status for the cluster entry with quorate = 1 and tolerates transient API errors while corosync and pmxcfs restart.

The wait is bounded and throws TimeoutException on expiry. It follows the -Wait timeout convention already used by Stop-PveContainer, Reset-PveVm and New-PveBackup: [ValidateRange(1, 3600)] public int Timeout with a default (60 here), no 0 = infinite.

Note there are two distinct timeout conventions in this module; do not mix them:

  • -Wait waits (Timeout, int with a default, range 1-3600, no infinite) — task/state waits.
  • HTTP client timeouts (TimeoutSeconds, int?, range 0-int.MaxValue, 0 = infinite) — Connect-PveServer, Send-PveFile, Invoke-PveStorageDownload, which set HttpClient.Timeout.

A single-node cluster reaches quorum in seconds (~6 s observed), so a node still not quorate after 60 s is broken rather than slow.

-Wait on every other cmdlet still means "wait for the task". Cluster creation is the exception because the task completing does not make the cluster usable.

Rationale

PVE's cluster-create task returns before corosync converges. Until the node is quorate it rejects a join with cluster not ready - no quorum?, so the natural sequence New-PveCluster -WaitAdd-PveClusterMember fails intermittently for every caller.

Observed on node A in integration run 172: the create task returned, corosync started ~1 s later, and node has quorum appeared ~6 s after that. The integration test had guarded this with Start-Sleep -Seconds 5 — a fixed sleep against a longer, variable convergence — which is why the cluster tests had never passed.

Anti-pattern (do not reintroduce)

# NEVER guard cluster convergence with a fixed sleep
New-PveCluster -ClusterName 'c1' -Wait
Start-Sleep -Seconds 5
Add-PveClusterMember ...

Correct pattern

# -Wait already guarantees quorum; join immediately
New-PveCluster -ClusterName 'c1' -Wait
Add-PveClusterMember ...

D015 — Lifecycle -Wait blocks until the guest config lock clears

Status: Superseded by D016 (2026-09-01) — the mechanism below is wrong Finding refs: (none — found via integration runs 183/184, 2026-09-01) Resolved in scan: n/a

This entry misdiagnosed the failure it was written for. Two different things in PVE are called "lock": the config lock (the lock: property — migrate, backup, clone, snapshot — a persisted config field, exposed as lock in status/current) and the flock on /var/lock/qemu-server/lock-<vmid>.conf taken by PVE::QemuConfig->lock_config, which is not exposed through the API at all. The integration failures were the flock; this entry guards the config lock, which an ordinary start/stop never sets. Run 186 confirmed the check never fired — Restart-PveVm took 4.11 s, unchanged. The real cause and fix are in D016. The waiting behaviour described below is harmless and still applies when a genuine config lock is present, so the code stays.

Decision

WaitForStatusTransition returns only when the guest reports the expected status and its config lock has cleared. Reaching the status is not enough: PVE publishes the new status while the operation still holds /var/lock/qemu-server/lock-<vmid>.conf, and the next API call against that guest fails with got timeout trying to take the same lock.

lock is read from the status/current response the poll already fetches — it is present on both qemu and lxc status/current and has been since PVE 5.4, well below this module's 7.0 floor, so this costs no extra request.

If the guest still reports the expected status on the final poll but the lock outlasts -Timeout, the cmdlet returns success rather than throwing. The waited-for operation did complete; only the settling ran long. This keeps a call that succeeded before the change from becoming an exception after it.

That fallback tests the most recent observation, not "matched at some point during the wait". A guest that reached the expected status and then drifted away from it has not satisfied the wait and still raises PveTaskTimeoutException. A poll that fails outright leaves the previous observation standing, so a single API blip is not read as divergence.

This is the same family as D014: a PVE task completing does not mean the resource is ready for the next operation. D014 is the cluster-quorum instance, D015 the guest-lock instance.

Rationale

Integration run 183 failed four tests from one cause. Restart-PveVm -Wait returned after 4.1 s having observed running; the following Stop-PveVm spent exactly 10.0 s failing to acquire the lock, which cascaded into the template convert, clone, and remove tests. Run 184, the same commit re-run, passed: its status poll happened to take 10.1 s, by which point the lock had cleared. The same settling happens either way — the only variable is whether the wait absorbs it or the next caller does.

The check lives in WaitForStatusTransition rather than in each cmdlet because all nine lifecycle call sites route through it.

Anti-pattern (do not reintroduce)

// NEVER treat the status transition alone as "ready for the next operation"
if (string.Equals(effectiveStatus, expectedStatus, StringComparison.OrdinalIgnoreCase))
    return task;

Correct pattern

var snapshot = GuestStatusSnapshot.Evaluate(json, expectedStatus);
if (snapshot.StatusMatched && !snapshot.Locked)
    return task;

D016 — Restart-PveVm uses PVE's native reboot endpoint

Status: Active Finding refs: (none — found via integration runs 183/185/186, root-caused on a live PVE 9.2.2 node 2026-09-01) Resolved in scan: n/a

Decision

Restart-PveVm calls POST /nodes/{node}/qemu/{vmid}/status/reboot (VmService.RebootVm). It must not compose a restart client-side as status/shutdown followed by status/start.

Rationale

Composing the restart races Proxmox's own post-stop cleanup for the guest's config flock:

  1. The shutdown completes and the QEMU process exits.
  2. qmeventd forks /usr/sbin/qm cleanup <vmid> ....
  3. The client sees status == stopped and immediately posts status/start. vm_start takes the flock, wins the race, starts a new QEMU, releases.
  4. qm cleanup then takes the flock with a 60 s timeout and polls vm_running_locally for up to 30 s, holding it the whole time, because it sees the new PID as the old one failing to exit. PVE's own warning names this: "QEMU process $pid for VM $vmid still running (or newly started)".
  5. Every subsequent call fails: lock_config defaults to 10 s, so the client gets can't lock file '/var/lock/qemu-server/lock-<vmid>.conf' - got timeout.

Measured on a reproduction (integration run 187), three distinct source constants matching:

qmstart    ends  t+3    <- qm cleanup takes the flock, sees the NEW pid
qmstop #1  FAIL  t+14   10 s = lock_config default
qmstop #2  FAIL  t+24   10 s = lock_config default
qmclone    FAIL  t+25    1 s = qmclone's separate source-VM lock timeout
qmstop #3  OK    t+33   <- released; hold was t+3..t+33 = 30 s = cleanup's wait loop

vm_reboot avoids all of it by holding the config lock across the entire shutdown and letting qm cleanup perform the restart while it already holds that same lock — there is no window for a client call to interleave.

This surfaced on PVE 9.2 and not 9.1 because of two May 2026 qemu-server changes (cleanup deduplication, shipped for 9.1.13, and the 30 s cleanup wait). Neither touches the REST surface, so the API changelog showed nothing — "the API did not change, therefore behaviour did not" is not a valid inference for this class of bug.

Containers are not affected by this decision: /nodes/{node}/lxc/{vmid}/status/reboot does not exist, so Restart-PveContainer necessarily keeps shutdown + start.

Anti-pattern (do not reintroduce)

// NEVER compose a VM restart from two client calls — it races qmeventd's cleanup
var shutdownTask = vmService.ShutdownVm(session, node, vmid, timeout);
WaitForStatusTransition(session, node, shutdownTask, vmid, "stopped", timeout);
var startTask = vmService.StartVm(session, node, vmid);

Correct pattern

var task = vmService.RebootVm(session, node, vmid, timeout);
if (Wait.IsPresent)
    task = WaitForStatusTransition(session, node, task, vmid, "running", timeout);

D017 — CI runs two lanes: a pinned gating lane and a report-only currency lane

Status: Active Finding refs: (none — arose from integration runs 176180, root-caused 2026-09-01) Resolved in scan: n/a

Decision

CI provisions nested PVE nodes in two distinct modes, and they must not be merged into one:

  • Lane 1, integration-tests.yml — nodes stay pinned to what the ISO ships. first-boot.sh must never run apt-get upgrade or dist-upgrade. This lane gates merges.
  • Lane 2, package-currency.yml — nodes are dist-upgraded to current PVE and the suite runs against them. Report-only: test failures do not fail the job.

Both declare concurrency: group: integration-tests. They drive the same nested VMIDs on the same parent node, so they must never run at once.

Rationale

apt-get upgrade holds back packages that need new dependencies. On the nested nodes that produced pve-cluster 9.1.6 against libpve-cluster-api-perl 9.1.0 — a combination no real install ever has — and its symptom was not a package error. The node's pmxcfs came back in local mode after a cluster join, /etc/pve/corosync.conf never appeared, and the node reported online=0 while corosync itself had healthy 2-node membership. Three CI runs went into diagnosing that, and removing the upgrade was the entire fix: run 180 was the first fully green integration run.

So the pin is what makes lane 1 a trustworthy merge gate. But a permanently pinned CI never exercises the module against a current PVE, and that gap is exactly where an upstream regression would hide. Lane 2 closes it without putting the instability back into the gating path.

Report-only is deliberate. A scheduled job that goes red on an upstream change nobody has chosen to chase becomes noise, and a noisy cron gets ignored — which is the failure mode that makes a canary worthless. The signal is the rolling issue and the recorded package set, not the check colour.

Consequence accepted: a module genuinely broken against current PVE shows a green weekly check plus an updated issue. Operator ruling 2026-09-01, to be revisited after a few releases.

A failure of the lane's own machinery — provisioning, the upgrade, the reboot, an unreachable node — still fails the job. run-integration.sh returns 3 for a genuine test failure and 4 when it cannot reach or authenticate to a node; only 3 is suppressed. Suppressing both would let a botched reboot report success while the lane learned nothing.

Amendment 2026-09-01 — the pin covers the test tooling, not just the nested PVE packages

The gating lane's own tooling is pinned by exact version for the same reason its nested PVE packages are: Pester in tests/Dockerfile.test (ARG PESTER_VERSION) and in .github/workflows/unit-tests.yml (env.PESTER_VERSION), installed and imported with -RequiredVersion at every site, including the suite's own import inside the container. The Dockerfile promotes the ARG to ENV so the version is discoverable at runtime. Both files must name the same version, and shell-selfchecks asserts it.

Before this, both sites used -MinimumVersion 5.0 with no ceiling. The image is rebuilt on every CI run and Pester is installed fresh on every unit-test run, so PSGallery decided the version — a new major could reach the merge gate with no commit to this repository, surfacing as unexplained test breakage on whichever PR happened to run next. That is the same class of moving input the lane split exists to eliminate; the difference is only that it moves in the test runner rather than in PVE.

It had already happened silently: steps named "Install Pester 5" were resolving 6.1.0 on both the PowerShell 5.1 and 7.x legs, because Pester 6 declares PowerShellVersion 5.1 and so installs on Windows PowerShell too. Nothing broke — the suite uses only constructs common to 5 and 6 — but nobody chose it.

Bumping is a deliberate commit that changes both files together.

Anti-pattern (do not reintroduce)

# NEVER in first-boot.sh — this is the mismatch that left a node unclustered
apt-get update -qq
apt-get -y upgrade

Correct pattern

# first-boot.sh installs only what provisioning needs; the ISO is the pin
apt-get update -qq
apt-get install -y -qq --no-install-recommends qemu-guest-agent open-iscsi
# package-currency.yml opts in explicitly; lane 1 never sets this
env:
  PVE_DIST_UPGRADE: '1'

D018 — The currency lane reboots after dist-upgrade, and proves it rebooted

Status: Active Finding refs: (none — found in pre-push review of PR #106, 2026-09-01) Resolved in scan: n/a

Decision

After dist-upgrade, prepare-test-environment.sh reboots the node unconditionally and then verifies the reboot happened by comparing /proc/sys/kernel/random/boot_id before and after. An unchanged boot id is fatal. The reboot lives here, not in first-boot.sh.

Rationale

A PVE dist-upgrade pulls proxmox-kernel-*. Without a reboot the node runs new userspace on the old kernel, so the lane records a package set it never actually ran and is blind to kernel regressions — it would report "current PVE" while testing something that never booted.

The reboot cannot live in first-boot.sh: that runs ordering = "fully-up" while the parent is still polling, so wait-for-pve.sh can discover the IP, see the API, pass auth, and then have the node reboot out from under provisioning. That presents as an intermittent network fault.

It is unconditional rather than gated on /var/run/reboot-required, because that file comes from update-notifier-common, which is not guaranteed present on a PVE node.

Verification is the part that is easy to omit and was omitted in the first draft. ssh … reboot returns non-zero when the connection dies, so it needs || true — which swallows every ssh failure, including the reboot never being issued. wait-for-api.sh then matches the still-running pre-reboot pveproxy on its first poll and returns responsive after 0s. The script exits 0 having proved nothing. A blind sleep before polling does not fix this; it is wrong in both directions and verifies nothing either way.

Order matters: prove the boot id changed first (ssh returns before pveproxy does), then wait for the API, then wait for pmxcfs — pvesm set writes /etc/pve/storage.cfg, which needs /etc/pve mounted, and on a fresh boot that lags the API by seconds.

Anti-pattern (do not reintroduce)

# NEVER — `|| true` hides a reboot that never happened, and the poll then
# matches the pre-reboot node and returns immediately
${SSH_CMD} "systemctl reboot" || true
sleep 30
bash "${SCRIPT_DIR}/wait-for-api.sh" "${NESTED_IP}" 8006 600

Correct pattern

boot_before="$(${SSH_CMD} "cat /proc/sys/kernel/random/boot_id")"
${SSH_CMD} "systemctl reboot" || true
boot_after=""
for _ in $(seq 1 60); do
    boot_after="$(${SSH_CMD} "cat /proc/sys/kernel/random/boot_id" 2>/dev/null || true)"
    [[ -n "${boot_after}" && "${boot_after}" != "${boot_before}" ]] && break
    sleep 5
done
if [[ -z "${boot_after}" || "${boot_after}" == "${boot_before}" ]]; then
    echo "ERROR: ${NESTED_IP} did not reboot (boot_id unchanged)" >&2
    exit 1
fi
bash "${SCRIPT_DIR}/wait-for-api.sh" "${NESTED_IP}" 8006 600

D019 — Local dev calls run-integration.sh directly; there is no wrapper script

Status: Active Finding refs: (none — found auditing the local dev path against the post-ARC CI, 2026-09-01) Resolved in scan: n/a

Decision

tests/infrastructure/scripts/run-integration.sh is the only entry point to the provision → test → cleanup lifecycle, for CI and for local development alike. Local runs invoke it inside the dev-infra container — the same image CI runs its jobs in. Do not add a convenience wrapper around it.

Build and unit tests need no container at all; they run natively against the solution.

Rationale

tests/dev.ps1 was a 291-line PowerShell wrapper over roughly six docker compose and docker exec calls. Every capability it had was already available elsewhere: build and unit tests are plain dotnet and Invoke-Pester invocations, and the module build it performed is duplicated inside run-integration.sh itself, which publishes and installs the module before running the suite.

Being a second entry point, it drifted from the script it wrapped and from the CI it claimed to replicate. By the time it was removed it still offered a -Version 8 leg retired in #88, mounted the Docker socket for storage containers replaced by the storage VM in #87, and defaulted its remote-host examples to a runner decommissioned in the ARC migration.

Four documentation files described a positional calling convention (./tests/dev.ps1 test) that did not do what it read as. The script took its actions from switches (-Test), but also declared [string[]] $Tests, so the bare word bound to -Tests — the integration-area filter. With no action switch set, the script then fell through to its -Shell default and silently opened an interactive container shell. Every documented command was wrong, and wrong in the quietest possible way: it succeeded at something nobody asked for.

A wrapper that must be kept in sync with the thing it wraps earns its place only when it removes real friction. This one removed none.

Anti-pattern (do not reintroduce)

A dev.ps1, Makefile target, or shell function that re-implements provisioning steps, module installation, or test invocation. If a local flow is awkward, fix it in run-integration.sh so CI gets the fix too.