add rack-topology package feature

This adds the rack-topology package feature with possible values of
single-sled or multi-sled. single-sled is intended for dev/CI
deployments, while multi-sled is intended for dogfood/prod. The value
of this determines which nexus config-partial.toml is packaged.

Right now the only difference single/multi is the crucible region
allocation strategy.

Author: faithanalog

.github/buildomat/jobs/package.sh

@@ -45,7 +45,7 @@ ptime -m ./tools/ci_download_softnpu_machinery
 
 # Build the test target
 ptime -m cargo run --locked --release --bin omicron-package -- \
-  -t test target create -i standard -m non-gimlet -s softnpu
+  -t test target create -i standard -m non-gimlet -s softnpu -r single-sled
 ptime -m cargo run --locked --release --bin omicron-package -- \
   -t test package
 
@@ -83,7 +83,7 @@ stamp_packages() {
 
 # Build necessary for the global zone
 ptime -m cargo run --locked --release --bin omicron-package -- \
-  -t host target create -i standard -m gimlet -s asic
+  -t host target create -i standard -m gimlet -s asic -r multi-sled
 ptime -m cargo run --locked --release --bin omicron-package -- \
   -t host package
 stamp_packages omicron-sled-agent maghemite propolis-server overlay

.github/buildomat/jobs/tuf-repo.sh

@@ -75,7 +75,7 @@ done
 mkdir /work/package
 pushd /work/package
 tar xf /input/package/work/package.tar.gz out package-manifest.toml target/release/omicron-package
-target/release/omicron-package -t default target create -i standard -m gimlet -s asic
+target/release/omicron-package -t default target create -i standard -m gimlet -s asic -r multi-sled
 ln -s /input/package/work/zones/* out/
 rm out/switch-softnpu.tar.gz  # not used when target switch=asic
 rm out/omicron-gateway-softnpu.tar.gz  # not used when target switch=asic

docs/how-to-run.adoc

@@ -287,20 +287,42 @@ Error: Creates a new build target, and sets it as "active"
 Usage: omicron-package target create [OPTIONS]
 
 Options:
-  -i, --image <IMAGE>      [default: standard] [possible values: standard, trampoline]
-  -m, --machine <MACHINE>  [possible values: gimlet, gimlet-standalone, non-gimlet]
-  -s, --switch <SWITCH>    [possible values: asic, stub, softnpu]
-  -h, --help               Print help (see more with '--help')
+  -i, --image <IMAGE>
+          [default: standard]
+
+          Possible values:
+          - standard:   A typical host OS image
+          - trampoline: A recovery host OS image, intended to bootstrap a Standard image
+
+  -m, --machine <MACHINE>
+          Possible values:
+          - gimlet:            Use sled agent configuration for a Gimlet
+          - gimlet-standalone: Use sled agent configuration for a Gimlet running in isolation
+          - non-gimlet:        Use sled agent configuration for a device emulating a Gimlet
+
+  -s, --switch <SWITCH>
+          Possible values:
+          - asic:    Use the "real" Dendrite, that attempts to interact with the Tofino
+          - stub:    Use a "stub" Dendrite that does not require any real hardware
+          - softnpu: Use a "softnpu" Dendrite that uses the SoftNPU asic emulator
+
+  -r, --rack-topology <RACK_TOPOLOGY>
+          Possible values:
+          - multi-sled:  Use configurations suitable for a multi-sled deployment, such as dogfood and production racks
+          - single-sled: Use configurations suitable for a single-sled deployment, such as CI and dev machines
+
+  -h, --help
+          Print help (see a summary with '-h')
 
 ----
 
 To setup a build target for a non-Gimlet machine with simulated (but fully functional) external networking, you would run:
 
 [source,console]
 ----
-$ cargo run --release --bin omicron-package -- -t default target create -i standard -m non-gimlet -s softnpu
+$ cargo run --release --bin omicron-package -- -t default target create -i standard -m non-gimlet -s softnpu -r single-sled
     Finished release [optimized] target(s) in 0.66s
-     Running `target/release/omicron-package -t default target create -i standard -m non-gimlet -s softnpu`
+     Running `target/release/omicron-package -t default target create -i standard -m non-gimlet -s softnpu -r single-sled`
 Created new build target 'default' and set it as active
 ----
 

package-manifest.toml

@@ -92,7 +92,7 @@ source.rust.binary_names = ["nexus", "schema-updater"]
 source.rust.release = true
 source.paths = [
   { from = "/opt/ooce/pgsql-13/lib/amd64", to = "/opt/ooce/pgsql-13/lib/amd64" },
-  { from = "smf/nexus", to = "/var/svc/manifest/site/nexus" },
+  { from = "smf/nexus/{{rack-topology}}", to = "/var/svc/manifest/site/nexus" },
   { from = "out/console-assets", to = "/var/nexus/static" },
   { from = "schema/crdb", to = "/var/nexus/schema/crdb" },
 ]

package/src/bin/omicron-package.rs

@@ -187,11 +187,12 @@ async fn do_target(
     let target_dir = artifact_dir.join("target");
     tokio::fs::create_dir_all(&target_dir).await?;
     match subcommand {
-        TargetCommand::Create { image, machine, switch } => {
+        TargetCommand::Create { image, machine, switch, rack_topology } => {
             let target = KnownTarget::new(
                 image.clone(),
                 machine.clone(),
                 switch.clone(),
+                rack_topology.clone(),
             )?;
 
             let path = get_single_target(&target_dir, name).await?;

package/src/lib.rs

@@ -46,6 +46,13 @@ pub enum TargetCommand {
 
         #[clap(short, long, default_value_if("image", "standard", "stub"))]
         switch: Option<crate::target::Switch>,
+
+        #[clap(
+            short,
+            long,
+            default_value_if("image", "standard", "single-sled")
+        )]
+        rack_topology: Option<crate::target::RackTopology>,
     },
     /// List all existing targets
     List,

package/src/target.rs

@@ -48,19 +48,35 @@ pub enum Switch {
     SoftNpu,
 }
 
+/// Topology of the sleds within the rack.
+#[derive(Clone, Debug, strum::EnumString, strum::Display, ValueEnum)]
+#[strum(serialize_all = "kebab-case")]
+#[clap(rename_all = "kebab-case")]
+pub enum RackTopology {
+    /// Use configurations suitable for a multi-sled deployment, such as dogfood
+    /// and production racks.
+    MultiSled,
+
+    /// Use configurations suitable for a single-sled deployment, such as CI and
+    /// dev machines.
+    SingleSled,
+}
+
 /// A strongly-typed variant of [Target].
 #[derive(Clone, Debug)]
 pub struct KnownTarget {
     image: Image,
     machine: Option<Machine>,
     switch: Option<Switch>,
+    rack_topology: Option<RackTopology>,
 }
 
 impl KnownTarget {
     pub fn new(
         image: Image,
         machine: Option<Machine>,
         switch: Option<Switch>,
+        rack_topology: Option<RackTopology>,
     ) -> Result<Self> {
         if matches!(image, Image::Trampoline) {
             if machine.is_some() {
@@ -77,7 +93,7 @@ impl KnownTarget {
             bail!("'switch=asic' is only valid with 'machine=gimlet'");
         }
 
-        Ok(Self { image, machine, switch })
+        Ok(Self { image, machine, switch, rack_topology })
     }
 }
 
@@ -87,6 +103,7 @@ impl Default for KnownTarget {
             image: Image::Standard,
             machine: Some(Machine::NonGimlet),
             switch: Some(Switch::Stub),
+            rack_topology: Some(RackTopology::MultiSled),
         }
     }
 }
@@ -101,6 +118,9 @@ impl From<KnownTarget> for Target {
         if let Some(switch) = kt.switch {
             map.insert("switch".to_string(), switch.to_string());
         }
+        if let Some(rack_topology) = kt.rack_topology {
+            map.insert("rack-topology".to_string(), rack_topology.to_string());
+        }
         Target(map)
     }
 }
@@ -121,6 +141,7 @@ impl std::str::FromStr for KnownTarget {
         let mut image = Self::default().image;
         let mut machine = None;
         let mut switch = None;
+        let mut rack_topology = None;
 
         for (k, v) in target.0.into_iter() {
             match k.as_str() {
@@ -133,6 +154,9 @@ impl std::str::FromStr for KnownTarget {
                 "switch" => {
                     switch = Some(v.parse()?);
                 }
+                "rack-topology" => {
+                    rack_topology = Some(v.parse()?);
+                }
                 _ => {
                     bail!(
                         "Unknown target key {k}\nValid keys include: [{}]",
@@ -146,6 +170,6 @@ impl std::str::FromStr for KnownTarget {
                 }
             }
         }
-        KnownTarget::new(image, machine, switch)
+        KnownTarget::new(image, machine, switch, rack_topology)
     }
 }

sled-agent/Cargo.toml

@@ -119,3 +119,5 @@ machine-non-gimlet = []
 switch-asic = []
 switch-stub = []
 switch-softnpu = []
+rack-topology-single-sled = []
+rack-topology-multi-sled = []

smf/nexus/config-partial.toml renamed to smf/nexus/multi-sled/config-partial.toml

@@ -0,0 +1,45 @@
+#
+# Oxide API: partial configuration file
+#
+
+[console]
+# Directory for static assets. Absolute path or relative to CWD.
+static_dir = "/var/nexus/static"
+session_idle_timeout_minutes = 60
+session_absolute_timeout_minutes = 480
+
+[authn]
+schemes_external = ["session_cookie", "access_token"]
+
+[log]
+# Show log messages of this level and more severe
+level = "debug"
+mode = "file"
+path = "/dev/stdout"
+if_exists = "append"
+
+# TODO: Uncomment the following lines to enable automatic schema
+# migration on boot.
+#
+# [schema]
+# schema_dir = "/var/nexus/schema/crdb"
+
+[background_tasks]
+dns_internal.period_secs_config = 60
+dns_internal.period_secs_servers = 60
+dns_internal.period_secs_propagation = 60
+dns_internal.max_concurrent_server_updates = 5
+dns_external.period_secs_config = 60
+dns_external.period_secs_servers = 60
+dns_external.period_secs_propagation = 60
+dns_external.max_concurrent_server_updates = 5
+# How frequently we check the list of stored TLS certificates.  This is
+# approximately an upper bound on how soon after updating the list of
+# certificates it will take _other_ Nexus instances to notice and stop serving
+# them (on a sunny day).
+external_endpoints.period_secs = 60
+
+[default_region_allocation_strategy]
+# by default, allocate without requirement for distinct sleds.
+# seed is omitted so a new seed will be chosen with every allocation.
+type = "random"