diff --git a/crates/ecstore/src/api/mod.rs b/crates/ecstore/src/api/mod.rs
index 672e290dc..9b7987387 100644
--- a/crates/ecstore/src/api/mod.rs
+++ b/crates/ecstore/src/api/mod.rs
@@ -326,6 +326,7 @@ pub mod global {
 }
 
 pub mod runtime {
+    pub use crate::runtime::instance::{InstanceContext, bootstrap_ctx};
     pub use crate::runtime::sources::{
         boot_time, bucket_monitor, deployment_id, endpoint_pools, expiry_state_handle, first_cluster_node_is_local,
         global_lock_client, global_lock_clients, global_tier_config_mgr, local_disk_map_read, object_store_handle, region,
@@ -389,8 +390,9 @@ pub mod store_list {
 pub mod storage {
     pub use crate::store::HealWalkVersion;
     pub use crate::store::{
-        ECStore, all_local_disk, all_local_disk_path, find_local_disk_by_ref, init_local_disks, init_lock_clients,
-        prewarm_local_disk_id_map,
+        ECStore, all_local_disk, all_local_disk_path, find_local_disk_by_ref, init_local_disks,
+        init_local_disks_with_instance_ctx, init_lock_clients, prewarm_local_disk_id_map,
+        prewarm_local_disk_id_map_with_instance_ctx,
     };
 }
 
diff --git a/crates/ecstore/src/core/sets.rs b/crates/ecstore/src/core/sets.rs
index 16b6f183b..425f05ec5 100644
--- a/crates/ecstore/src/core/sets.rs
+++ b/crates/ecstore/src/core/sets.rs
@@ -102,6 +102,21 @@ impl Sets {
         fm: &FormatV3,
         pool_idx: usize,
         parity_count: usize,
+    ) -> Result<Arc<Self>> {
+        Self::new_with_instance_ctx(disks, endpoints, fm, pool_idx, parity_count, bootstrap_ctx()).await
+    }
+
+    /// Build the pool's sets bound to an explicit instance context (Phase 5
+    /// follow-up, backlog#1052). The legacy [`Sets::new`] entry adopts the
+    /// process bootstrap context; a store constructed around its own context
+    /// passes it here so the whole object graph shares one cell.
+    pub async fn new_with_instance_ctx(
+        disks: Vec<Option<DiskStore>>,
+        endpoints: &PoolEndpoints,
+        fm: &FormatV3,
+        pool_idx: usize,
+        parity_count: usize,
+        instance_ctx: Arc<InstanceContext>,
     ) -> Result<Arc<Self>> {
         let set_count = fm.erasure.sets.len();
         let set_drive_count = fm.erasure.sets[0].len();
@@ -127,8 +142,8 @@ impl Sets {
                     continue;
                 }
 
-                if disk.as_ref().unwrap().is_local() && runtime_sources::setup_is_dist_erasure().await {
-                    let local_disk = runtime_sources::local_disk_set_drive(pool_idx, i, j).await;
+                if disk.as_ref().unwrap().is_local() && instance_ctx.is_dist_erasure().await {
+                    let local_disk = runtime_sources::local_disk_set_drive(&instance_ctx, pool_idx, i, j).await;
 
                     if local_disk.is_none() {
                         warn!("sets new set_drive {}-{} local_disk is none", i, j);
@@ -163,7 +178,7 @@ impl Sets {
                 .as_ref()
                 .map(|registry| registry.clients_for_endpoints(&set_endpoints))
                 .unwrap_or_default();
-            let set_disks = SetDisks::new(
+            let set_disks = SetDisks::new_with_instance_ctx(
                 runtime_sources::local_node_name().await,
                 Arc::new(RwLock::new(set_drive)),
                 set_drive_count,
@@ -173,6 +188,7 @@ impl Sets {
                 set_endpoints,
                 fm.clone(),
                 lockers,
+                instance_ctx.clone(),
             )
             .await;
 
@@ -193,10 +209,7 @@ impl Sets {
             default_parity_count: parity_count,
             distribution_algo: fm.erasure.distribution_algo.clone(),
             exit_signal: Some(tx),
-            // Single-instance: same bootstrap context the owning ECStore adopts
-            // (constructed before the store, so sourced here directly). Slice 8
-            // threads a per-instance context in for true multi-instance.
-            ctx: bootstrap_ctx(),
+            ctx: instance_ctx,
         });
 
         let asets = sets.clone();
@@ -1462,4 +1475,65 @@ mod tests {
             "unformatted disk must be Missing on its real index, not on a placeholder"
         );
     }
+
+    fn instance_ctx_test_pool_endpoints() -> (FormatV3, PoolEndpoints) {
+        let format = FormatV3::new(1, 2);
+        let endpoints = vec![
+            Endpoint::try_from("http://127.0.0.1:9000/data0").expect("first endpoint should parse"),
+            Endpoint::try_from("http://127.0.0.1:9001/data1").expect("second endpoint should parse"),
+        ];
+        let pool_endpoints = PoolEndpoints {
+            legacy: false,
+            set_count: 1,
+            drives_per_set: 2,
+            endpoints: Endpoints::from(endpoints),
+            cmd_line: "instance-ctx-adoption-test".to_string(),
+            platform: "test".to_string(),
+        };
+        (format, pool_endpoints)
+    }
+
+    // Phase 5 follow-up (backlog#1052): a pool built through the ctx-explicit
+    // constructor carries the caller's context through Sets AND every SetDisks,
+    // so nothing in the object graph silently binds to the process bootstrap.
+    #[tokio::test]
+    async fn sets_new_with_instance_ctx_threads_context_through_graph() {
+        let (format, pool_endpoints) = instance_ctx_test_pool_endpoints();
+        let instance_ctx = Arc::new(InstanceContext::new());
+
+        let sets = Sets::new_with_instance_ctx(vec![None, None], &pool_endpoints, &format, 0, 1, instance_ctx.clone())
+            .await
+            .expect("sets should build with empty disks");
+
+        assert!(
+            Arc::ptr_eq(sets.instance_ctx(), &instance_ctx),
+            "Sets must adopt the explicitly passed instance context"
+        );
+        for set_disks in &sets.disk_set {
+            assert!(
+                Arc::ptr_eq(set_disks.instance_ctx(), &instance_ctx),
+                "every SetDisks must adopt the explicitly passed instance context"
+            );
+        }
+        assert!(
+            !Arc::ptr_eq(sets.instance_ctx(), &bootstrap_ctx()),
+            "a fresh context must not alias the process bootstrap context"
+        );
+    }
+
+    // The legacy constructor keeps single-instance behavior byte-for-byte: it
+    // still adopts the process bootstrap context.
+    #[tokio::test]
+    async fn sets_new_legacy_adopts_bootstrap_context() {
+        let (format, pool_endpoints) = instance_ctx_test_pool_endpoints();
+
+        let sets = Sets::new(vec![None, None], &pool_endpoints, &format, 0, 1)
+            .await
+            .expect("sets should build with empty disks");
+
+        assert!(
+            Arc::ptr_eq(sets.instance_ctx(), &bootstrap_ctx()),
+            "legacy Sets::new must keep adopting the process bootstrap context"
+        );
+    }
 }
diff --git a/crates/ecstore/src/runtime/sources.rs b/crates/ecstore/src/runtime/sources.rs
index b249ad322..94902c6b0 100644
--- a/crates/ecstore/src/runtime/sources.rs
+++ b/crates/ecstore/src/runtime/sources.rs
@@ -20,6 +20,7 @@ use std::{
 
 use crate::bucket::bandwidth::monitor::Monitor;
 use crate::disk::endpoint::Endpoint;
+use crate::runtime::instance::InstanceContext;
 use crate::{
     bucket::lifecycle::bucket_lifecycle_ops::{ExpiryState, TransitionState},
     bucket::metadata_sys::{BucketMetadataSys, get_global_bucket_metadata_sys},
@@ -33,8 +34,8 @@ use crate::{
         TypeLocalDiskSetDrives, get_background_services_cancel_token, get_global_bucket_monitor, get_global_deployment_id,
         get_global_endpoints, get_global_endpoints_opt, get_global_lock_client, get_global_lock_clients, get_global_region,
         get_global_tier_config_mgr, global_rustfs_port, init_global_bucket_monitor, is_dist_erasure, is_erasure, is_erasure_sd,
-        is_first_cluster_node_local, resolve_object_store_handle, set_global_deployment_id, set_global_lock_client,
-        set_global_lock_clients, set_object_layer, update_erasure_type,
+        is_first_cluster_node_local, resolve_object_store_handle, set_global_lock_client, set_global_lock_clients,
+        set_object_layer, update_erasure_type,
     },
     services::batch_processor::{GlobalBatchProcessors, get_global_processors},
     services::event_notification::EventNotifier,
@@ -278,12 +279,6 @@ pub(crate) fn replication_runtime_initialized() -> bool {
     crate::runtime::global::current_ctx().replication_initialized()
 }
 
-pub(crate) fn ensure_deployment_id(deployment_id: Uuid) {
-    if get_global_deployment_id().is_none() {
-        set_global_deployment_id(deployment_id);
-    }
-}
-
 pub fn global_lock_client() -> Option<Arc<dyn LockClient>> {
     get_global_lock_client()
 }
@@ -419,8 +414,8 @@ pub(crate) async fn clear_local_disk_id_map_for_test() {
     local_disk_id_map_handle().write().await.clear();
 }
 
-pub(crate) async fn record_local_disk_id(disk_id: Uuid, endpoint: String) {
-    local_disk_id_map_handle().write().await.insert(disk_id, endpoint);
+pub(crate) async fn record_local_disk_id(instance_ctx: &Arc<InstanceContext>, disk_id: Uuid, endpoint: String) {
+    instance_ctx.local_disk_id_map().write().await.insert(disk_id, endpoint);
 }
 
 pub(crate) async fn replace_local_disk_id(previous: Option<Uuid>, current: Option<Uuid>, endpoint: String) {
@@ -434,8 +429,8 @@ pub(crate) async fn replace_local_disk_id(previous: Option<Uuid>, current: Optio
     }
 }
 
-pub(crate) async fn record_local_disks(disks: Vec<DiskStore>) {
-    let map = local_disk_map_handle();
+pub(crate) async fn record_local_disks(instance_ctx: &Arc<InstanceContext>, disks: Vec<DiskStore>) {
+    let map = instance_ctx.local_disk_map();
     let mut global_local_disk_map = map.write().await;
     for disk in disks {
         let path = disk.endpoint().to_string();
@@ -443,8 +438,13 @@ pub(crate) async fn record_local_disks(disks: Vec<DiskStore>) {
     }
 }
 
-pub(crate) async fn local_disk_set_drive(pool_idx: usize, set_idx: usize, disk_idx: usize) -> Option<DiskStore> {
-    local_disk_set_drives_handle().read().await[pool_idx][set_idx][disk_idx].clone()
+pub(crate) async fn local_disk_set_drive(
+    instance_ctx: &Arc<InstanceContext>,
+    pool_idx: usize,
+    set_idx: usize,
+    disk_idx: usize,
+) -> Option<DiskStore> {
+    instance_ctx.local_disk_set_drives().read().await[pool_idx][set_idx][disk_idx].clone()
 }
 
 pub(crate) async fn local_disk_for_endpoint(endpoint: &Endpoint) -> Option<DiskStore> {
@@ -488,8 +488,12 @@ pub(crate) async fn local_disk_entries() -> Vec<Option<DiskStore>> {
     local_disk_map_handle().read().await.values().cloned().collect()
 }
 
-pub(crate) async fn initialize_local_disk_maps(endpoint_pools: EndpointServerPools, opt: &DiskOption) -> Result<()> {
-    let set_drives = local_disk_set_drives_handle();
+pub(crate) async fn initialize_local_disk_maps(
+    instance_ctx: &Arc<InstanceContext>,
+    endpoint_pools: EndpointServerPools,
+    opt: &DiskOption,
+) -> Result<()> {
+    let set_drives = instance_ctx.local_disk_set_drives();
     let mut global_set_drives = set_drives.write().await;
     for pool_eps in endpoint_pools.as_ref().iter() {
         let mut set_count_drives = Vec::with_capacity(pool_eps.set_count);
@@ -500,7 +504,7 @@ pub(crate) async fn initialize_local_disk_maps(endpoint_pools: EndpointServerPoo
         global_set_drives.push(set_count_drives);
     }
 
-    let map = local_disk_map_handle();
+    let map = instance_ctx.local_disk_map();
     let mut global_local_disk_map = map.write().await;
 
     for pool_eps in endpoint_pools.as_ref().iter() {
diff --git a/crates/ecstore/src/set_disk/mod.rs b/crates/ecstore/src/set_disk/mod.rs
index 32011ad74..2e2daa570 100644
--- a/crates/ecstore/src/set_disk/mod.rs
+++ b/crates/ecstore/src/set_disk/mod.rs
@@ -1744,10 +1744,39 @@ impl SetDisks {
         format: FormatV3,
         lockers: Vec<Arc<dyn LockClient>>,
     ) -> Arc<Self> {
-        // Single-instance sources the process bootstrap context (the one the
-        // owning ECStore adopts). Slice 8 threads a per-instance context in for
-        // true multi-instance.
-        let ctx = bootstrap_ctx();
+        Self::new_with_instance_ctx(
+            locker_owner,
+            disks,
+            set_drive_count,
+            default_parity_count,
+            set_index,
+            pool_index,
+            set_endpoints,
+            format,
+            lockers,
+            bootstrap_ctx(),
+        )
+        .await
+    }
+
+    /// Build a set bound to an explicit instance context (Phase 5 follow-up,
+    /// backlog#1052). The legacy [`SetDisks::new`] entry adopts the process
+    /// bootstrap context; a store constructed around its own context threads it
+    /// down here so the whole object graph shares one cell.
+    #[allow(clippy::too_many_arguments)]
+    pub async fn new_with_instance_ctx(
+        locker_owner: String,
+        disks: Arc<RwLock<Vec<Option<DiskStore>>>>,
+        set_drive_count: usize,
+        default_parity_count: usize,
+        set_index: usize,
+        pool_index: usize,
+        set_endpoints: Vec<Endpoint>,
+        format: FormatV3,
+        lockers: Vec<Arc<dyn LockClient>>,
+        instance_ctx: Arc<InstanceContext>,
+    ) -> Arc<Self> {
+        let ctx = instance_ctx;
         Arc::new(SetDisks {
             locker_owner,
             disks,
diff --git a/crates/ecstore/src/store/init.rs b/crates/ecstore/src/store/init.rs
index cd8925f34..69ec8f479 100644
--- a/crates/ecstore/src/store/init.rs
+++ b/crates/ecstore/src/store/init.rs
@@ -15,6 +15,7 @@
 use super::*;
 use crate::core::pools::local_decommission_queue_prefix;
 use crate::error::is_err_decommission_running;
+use crate::runtime::instance::InstanceContext;
 use crate::runtime::sources as runtime_sources;
 use crate::storage_api_contracts::object::EcstoreObjectIO;
 use tracing::{debug, error, info, warn};
@@ -165,6 +166,24 @@ impl ECStore {
     #[allow(clippy::new_ret_no_self)]
     #[instrument(level = "debug", skip(endpoint_pools))]
     pub async fn new(address: SocketAddr, endpoint_pools: EndpointServerPools, ctx: CancellationToken) -> Result<Arc<Self>> {
+        Self::new_with_instance_ctx(address, endpoint_pools, ctx, crate::runtime::instance::bootstrap_ctx()).await
+    }
+
+    /// Build a store around an explicit instance context (Phase 5 follow-up,
+    /// backlog#1052). The legacy [`ECStore::new`] entry adopts the process
+    /// bootstrap context, keeping single-instance startup byte-for-byte
+    /// unchanged; a caller that owns its own context (a future second embedded
+    /// server) passes it here so every construction-time write — pool sets,
+    /// local-disk registry, deployment id — lands on that context instead of
+    /// the shared bootstrap one.
+    #[allow(clippy::new_ret_no_self)]
+    #[instrument(level = "debug", skip(endpoint_pools, instance_ctx))]
+    pub async fn new_with_instance_ctx(
+        address: SocketAddr,
+        endpoint_pools: EndpointServerPools,
+        ctx: CancellationToken,
+        instance_ctx: Arc<InstanceContext>,
+    ) -> Result<Arc<Self>> {
         // let layouts = DisksLayout::from_volumes(endpoints.as_slice())?;
 
         let mut deployment_id = None;
@@ -308,15 +327,16 @@ impl ECStore {
                 }
             }
 
-            let sets = Sets::new(disks.clone(), pool_eps, &fm, i, common_parity_drives).await?;
+            let sets =
+                Sets::new_with_instance_ctx(disks.clone(), pool_eps, &fm, i, common_parity_drives, instance_ctx.clone()).await?;
             pools.push(sets);
 
             disk_map.insert(i, disks);
         }
 
         // Replace the local disk
-        if !runtime_sources::setup_is_dist_erasure().await {
-            runtime_sources::record_local_disks(local_disks).await;
+        if !instance_ctx.is_dist_erasure().await {
+            runtime_sources::record_local_disks(&instance_ctx, local_disks).await;
         }
 
         let peer_sys = S3PeerSys::new(&endpoint_pools);
@@ -334,14 +354,17 @@ impl ECStore {
             decommission_cancelers,
             start_gate: Mutex::new(()),
             pool_meta_save_gate: Mutex::new(()),
-            // Adopt the process bootstrap context so startup writes (erasure
-            // type recorded before this point) and later reads share one cell.
-            ctx: crate::runtime::instance::bootstrap_ctx(),
+            // Adopt the caller's context (the process bootstrap one on the
+            // legacy path) so startup writes (erasure type recorded before
+            // this point) and later reads share one cell.
+            ctx: instance_ctx.clone(),
         });
 
-        // Only set it when the global deployment ID is not yet configured
-        if let Some(dep_id) = deployment_id {
-            runtime_sources::ensure_deployment_id(dep_id);
+        // Only set it when this instance's deployment ID is not yet configured
+        if let Some(dep_id) = deployment_id
+            && instance_ctx.deployment_id().is_none()
+        {
+            instance_ctx.set_deployment_id(dep_id);
         }
 
         let wait_sec = 5;
@@ -759,4 +782,84 @@ mod tests {
             "the expanded pool should be initialized by its own first local endpoint"
         );
     }
+
+    // Phase 5 follow-up (backlog#1052): building a real store through the
+    // ctx-explicit constructor lands every construction-time write — object
+    // graph adoption, local-disk registry, deployment id — on the passed
+    // context, not on the process bootstrap one. This is the storage-layer
+    // seam a future second embedded server needs to stay isolated.
+    #[tokio::test]
+    async fn new_with_instance_ctx_threads_context_through_store_graph() {
+        use crate::runtime::instance::InstanceContext;
+
+        let temp_dir = tempfile::tempdir().expect("create temp store dir");
+        let disk_paths: Vec<_> = (1..=4).map(|i| temp_dir.path().join(format!("disk{i}"))).collect();
+        for path in &disk_paths {
+            tokio::fs::create_dir_all(path).await.expect("create disk dir");
+        }
+
+        let mut endpoints = Vec::new();
+        for (i, path) in disk_paths.iter().enumerate() {
+            let mut endpoint = Endpoint::try_from(path.to_str().expect("disk path should be utf-8")).expect("local endpoint");
+            endpoint.set_pool_index(0);
+            endpoint.set_set_index(0);
+            endpoint.set_disk_index(i);
+            endpoints.push(endpoint);
+        }
+        let endpoint_pools = EndpointServerPools(vec![PoolEndpoints {
+            legacy: false,
+            set_count: 1,
+            drives_per_set: 4,
+            endpoints: Endpoints::from(endpoints),
+            cmd_line: "instance-ctx-store-graph-test".to_string(),
+            platform: "test".to_string(),
+        }]);
+
+        let instance_ctx = Arc::new(InstanceContext::new());
+        crate::store::init_local_disks_with_instance_ctx(&instance_ctx, endpoint_pools.clone())
+            .await
+            .expect("register local disks into the fresh context");
+
+        let store = crate::store::ECStore::new_with_instance_ctx(
+            "127.0.0.1:0".parse().expect("test address"),
+            endpoint_pools,
+            CancellationToken::new(),
+            instance_ctx.clone(),
+        )
+        .await
+        .expect("store should build around the fresh context");
+
+        assert!(
+            Arc::ptr_eq(&store.ctx, &instance_ctx),
+            "the store must adopt the explicitly passed instance context"
+        );
+        for sets in &store.pools {
+            assert!(
+                Arc::ptr_eq(sets.instance_ctx(), &instance_ctx),
+                "every pool's Sets must carry the passed instance context"
+            );
+        }
+        assert_eq!(
+            instance_ctx.deployment_id(),
+            Some(store.id),
+            "the deployment id must land on the passed context and mirror the store id"
+        );
+
+        let registered: Vec<String> = instance_ctx.local_disk_map().read().await.keys().cloned().collect();
+        assert_eq!(registered.len(), 4, "the passed context must register all four local disks");
+        let bootstrap = crate::runtime::instance::bootstrap_ctx();
+        assert_ne!(
+            bootstrap.deployment_id(),
+            Some(store.id),
+            "the bootstrap context must not absorb the fresh store's deployment id"
+        );
+        let bootstrap_map = bootstrap.local_disk_map();
+        let bootstrap_map = bootstrap_map.read().await;
+        for key in &registered {
+            assert!(
+                !bootstrap_map.contains_key(key),
+                "the bootstrap context must not absorb the fresh store's disks"
+            );
+        }
+    }
 }
diff --git a/crates/ecstore/src/store/mod.rs b/crates/ecstore/src/store/mod.rs
index 3fa2b0dd2..f80859133 100644
--- a/crates/ecstore/src/store/mod.rs
+++ b/crates/ecstore/src/store/mod.rs
@@ -164,8 +164,9 @@ pub(crate) mod utils;
 
 use peer::init_local_peer;
 pub use peer::{
-    all_local_disk, all_local_disk_path, find_local_disk_by_ref, get_disk_infos, init_local_disks, init_lock_clients,
-    prewarm_local_disk_id_map,
+    all_local_disk, all_local_disk_path, find_local_disk_by_ref, get_disk_infos, init_local_disks,
+    init_local_disks_with_instance_ctx, init_lock_clients, prewarm_local_disk_id_map,
+    prewarm_local_disk_id_map_with_instance_ctx,
 };
 
 pub struct ECStore {
diff --git a/crates/ecstore/src/store/peer.rs b/crates/ecstore/src/store/peer.rs
index 21441efea..ac1062bff 100644
--- a/crates/ecstore/src/store/peer.rs
+++ b/crates/ecstore/src/store/peer.rs
@@ -13,6 +13,7 @@
 // limitations under the License.
 
 use super::*;
+use crate::runtime::instance::InstanceContext;
 use crate::runtime::sources as runtime_sources;
 use tracing::{debug, error};
 
@@ -22,8 +23,12 @@ const EVENT_LOCAL_DISK_ID_PREWARM_SKIPPED: &str = "local_disk_id_prewarm_skipped
 const EVENT_LOCK_CLIENT_INITIALIZATION_FAILED: &str = "lock_client_initialization_failed";
 
 async fn remember_local_disk_id(disk: &DiskStore) -> Option<Uuid> {
+    remember_local_disk_id_with_instance_ctx(&crate::runtime::global::current_ctx(), disk).await
+}
+
+async fn remember_local_disk_id_with_instance_ctx(instance_ctx: &Arc<InstanceContext>, disk: &DiskStore) -> Option<Uuid> {
     let disk_id = disk.get_disk_id().await.ok().flatten()?;
-    runtime_sources::record_local_disk_id(disk_id, disk.endpoint().to_string()).await;
+    runtime_sources::record_local_disk_id(instance_ctx, disk_id, disk.endpoint().to_string()).await;
     Some(disk_id)
 }
 
@@ -69,7 +74,21 @@ pub async fn all_local_disk() -> Vec<DiskStore> {
 }
 
 pub async fn prewarm_local_disk_id_map() {
-    for disk in all_local_disk().await {
+    prewarm_local_disk_id_map_with_instance_ctx(&crate::runtime::global::current_ctx()).await
+}
+
+/// Prewarm the disk-id map of an explicit instance context (Phase 5 follow-up,
+/// backlog#1052): startup passes the context whose disk map it just populated
+/// instead of resolving the process-level default.
+pub async fn prewarm_local_disk_id_map_with_instance_ctx(instance_ctx: &Arc<InstanceContext>) {
+    let disks: Vec<DiskStore> = instance_ctx
+        .local_disk_map()
+        .read()
+        .await
+        .values()
+        .filter_map(|v| v.as_ref().cloned())
+        .collect();
+    for disk in disks {
         if let Err(err) = disk.get_disk_id().await {
             debug!(
                 event = EVENT_LOCAL_DISK_ID_PREWARM_SKIPPED,
@@ -82,17 +101,28 @@ pub async fn prewarm_local_disk_id_map() {
             continue;
         }
 
-        let _ = remember_local_disk_id(&disk).await;
+        let _ = remember_local_disk_id_with_instance_ctx(instance_ctx, &disk).await;
     }
 }
 
 pub async fn init_local_disks(endpoint_pools: EndpointServerPools) -> Result<()> {
+    init_local_disks_with_instance_ctx(&crate::runtime::global::current_ctx(), endpoint_pools).await
+}
+
+/// Register the pools' local disks into an explicit instance context (Phase 5
+/// follow-up, backlog#1052). The legacy [`init_local_disks`] entry resolves the
+/// process-level default context; startup paths that own a context pass it here
+/// so a future second instance's disks cannot leak into the first one's registry.
+pub async fn init_local_disks_with_instance_ctx(
+    instance_ctx: &Arc<InstanceContext>,
+    endpoint_pools: EndpointServerPools,
+) -> Result<()> {
     let opt = &DiskOption {
         cleanup: true,
         health_check: true,
     };
 
-    runtime_sources::initialize_local_disk_maps(endpoint_pools, opt).await
+    runtime_sources::initialize_local_disk_maps(instance_ctx, endpoint_pools, opt).await
 }
 
 pub fn init_lock_clients(endpoint_pools: EndpointServerPools) {
@@ -182,3 +212,63 @@ pub async fn get_disk_infos(disks: &[Option<DiskStore>]) -> Vec<Option<DiskInfo>
 
     res
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::layout::endpoints::{Endpoints, PoolEndpoints};
+
+    fn single_local_disk_pools(dir: &std::path::Path) -> EndpointServerPools {
+        let mut endpoint = Endpoint::try_from(dir.to_str().expect("temp dir path should be utf-8")).expect("local endpoint");
+        endpoint.set_pool_index(0);
+        endpoint.set_set_index(0);
+        endpoint.set_disk_index(0);
+
+        EndpointServerPools(vec![PoolEndpoints {
+            legacy: false,
+            set_count: 1,
+            drives_per_set: 1,
+            endpoints: Endpoints::from(vec![endpoint]),
+            cmd_line: "instance-ctx-disk-registry-test".to_string(),
+            platform: "test".to_string(),
+        }])
+    }
+
+    // Phase 5 follow-up (backlog#1052): registering local disks through the
+    // ctx-explicit entry writes the passed context's registry only — the
+    // process bootstrap context (and any other instance) stays clean, so a
+    // future second server's disks cannot leak into the first one's registry.
+    #[tokio::test]
+    async fn init_local_disks_with_instance_ctx_isolates_disk_registry() {
+        let temp_dir = tempfile::tempdir().expect("create temp disk dir");
+        let endpoint_pools = single_local_disk_pools(temp_dir.path());
+        let instance_ctx = Arc::new(InstanceContext::new());
+
+        init_local_disks_with_instance_ctx(&instance_ctx, endpoint_pools)
+            .await
+            .expect("local disks should register into the passed context");
+
+        let registered: Vec<String> = instance_ctx.local_disk_map().read().await.keys().cloned().collect();
+        assert_eq!(registered.len(), 1, "the passed context must hold exactly the one local disk");
+        assert_eq!(
+            instance_ctx.local_disk_set_drives().read().await.len(),
+            1,
+            "the passed context must hold the pool/set/drive layout"
+        );
+
+        let bootstrap = crate::runtime::instance::bootstrap_ctx();
+        let bootstrap_map = bootstrap.local_disk_map();
+        let bootstrap_map = bootstrap_map.read().await;
+        let sibling = InstanceContext::new();
+        for key in &registered {
+            assert!(
+                !bootstrap_map.contains_key(key),
+                "bootstrap context must not absorb a disk registered into an explicit context"
+            );
+            assert!(
+                !sibling.local_disk_map().read().await.contains_key(key),
+                "a sibling context must not observe another instance's disks"
+            );
+        }
+    }
+}
diff --git a/rustfs/src/embedded.rs b/rustfs/src/embedded.rs
index bb2599c9b..fff5a6803 100644
--- a/rustfs/src/embedded.rs
+++ b/rustfs/src/embedded.rs
@@ -44,7 +44,8 @@
 //!
 //! Only **one `RustFSServer`** may exist per process because the underlying
 //! storage engine uses process-global singletons (`OnceLock`). Attempting to
-//! start a second server will return an error.
+//! start a second server will return an error. Removing this limitation is
+//! tracked in [backlog#1052](https://github.com/rustfs/backlog/issues/1052).
 
 use crate::server::ShutdownHandle;
 use crate::startup_embedded::{EmbeddedStartedServer, EmbeddedStartupArgs, EmbeddedStartupError, run_embedded_startup};
diff --git a/rustfs/src/startup_embedded.rs b/rustfs/src/startup_embedded.rs
index 20bac96c6..0a34481d0 100644
--- a/rustfs/src/startup_embedded.rs
+++ b/rustfs/src/startup_embedded.rs
@@ -24,6 +24,7 @@ use crate::{
     startup_services::init_embedded_startup_runtime_services,
     startup_shutdown::signal_embedded_startup_shutdown,
     startup_storage::{init_embedded_startup_storage_foundation, init_embedded_startup_storage_runtime},
+    storage_api::startup::storage::bootstrap_instance_ctx,
 };
 use std::{io, net::SocketAddr, path::PathBuf};
 use tokio_util::sync::CancellationToken;
@@ -108,6 +109,12 @@ pub(crate) async fn run_embedded_startup(args: EmbeddedStartupArgs) -> Result<Em
     // case callers can retry.
     let mut startup_guard = EmbeddedStartupGuard::new();
 
+    // The embedded server still shares the process bootstrap context (see the
+    // EMBEDDED_SERVER_STARTED guard); once the request path and app subsystems
+    // are per-server too (backlog#1052 S2–S4), each server constructs its own
+    // context here instead.
+    let instance_ctx = bootstrap_instance_ctx();
+
     let EmbeddedStartupConfig {
         config,
         identity,
@@ -120,13 +127,15 @@ pub(crate) async fn run_embedded_startup(args: EmbeddedStartupArgs) -> Result<Em
         .await
         .map_err(init_error)?;
 
-    let listen_context = init_embedded_startup_listen_context(&config).await.map_err(init_error)?;
+    let listen_context = init_embedded_startup_listen_context(&config, &instance_ctx)
+        .await
+        .map_err(init_error)?;
 
     startup_guard
         .mark_global_init_started()
         .map_err(|_| EmbeddedStartupError::AlreadyStarted)?;
 
-    let endpoint_pools = init_embedded_startup_storage_foundation(&listen_context.server_address, &config.volumes)
+    let endpoint_pools = init_embedded_startup_storage_foundation(&listen_context.server_address, &config.volumes, &instance_ctx)
         .await
         .map_err(init_error)?;
 
@@ -140,6 +149,7 @@ pub(crate) async fn run_embedded_startup(args: EmbeddedStartupArgs) -> Result<Em
         &endpoint_pools,
         listen_context.readiness.clone(),
         cancel_token.clone(),
+        instance_ctx,
     )
     .await
     {
diff --git a/rustfs/src/startup_entrypoint.rs b/rustfs/src/startup_entrypoint.rs
index 4d226ff97..ee46d522d 100644
--- a/rustfs/src/startup_entrypoint.rs
+++ b/rustfs/src/startup_entrypoint.rs
@@ -19,6 +19,7 @@ use crate::{
     startup_server::{StartupHttpServers, StartupListenContext, init_startup_http_servers, init_startup_listen_context},
     startup_services::init_startup_runtime_services,
     startup_storage::{StartupStorageRuntime, init_startup_storage_foundation, init_startup_storage_runtime},
+    storage_api::startup::storage::bootstrap_instance_ctx,
 };
 use std::io::{Error, Result};
 use tracing::{error, instrument};
@@ -103,13 +104,18 @@ async fn async_main() -> Result<()> {
 
 #[instrument(skip(config))]
 async fn run(config: Config) -> Result<()> {
+    // Single-instance startup threads the process bootstrap context through
+    // the storage path explicitly (Phase 5 follow-up, backlog#1052); a future
+    // multi-instance server constructs its own context here instead.
+    let instance_ctx = bootstrap_instance_ctx();
+
     let StartupListenContext {
         readiness,
         server_addr,
         server_address,
-    } = init_startup_listen_context(&config).await?;
+    } = init_startup_listen_context(&config, &instance_ctx).await?;
 
-    let endpoint_pools = init_startup_storage_foundation(&server_address, &config.volumes).await?;
+    let endpoint_pools = init_startup_storage_foundation(&server_address, &config.volumes, &instance_ctx).await?;
     let StartupHttpServers {
         state_manager,
         s3_shutdown_tx,
@@ -119,7 +125,7 @@ async fn run(config: Config) -> Result<()> {
     let StartupStorageRuntime {
         store,
         shutdown_token: ctx,
-    } = init_startup_storage_runtime(server_addr, &endpoint_pools, readiness.clone()).await?;
+    } = init_startup_storage_runtime(server_addr, &endpoint_pools, readiness.clone(), instance_ctx).await?;
 
     let service_runtime = init_startup_runtime_services(
         &config,
diff --git a/rustfs/src/startup_lifecycle.rs b/rustfs/src/startup_lifecycle.rs
index f66647509..869dd1e75 100644
--- a/rustfs/src/startup_lifecycle.rs
+++ b/rustfs/src/startup_lifecycle.rs
@@ -45,14 +45,17 @@ const EVENT_EMBEDDED_SERVER_STATE: &str = "embedded_server_state";
 ///
 /// Phase 5 (backlog#939) moved per-instance runtime state (erasure setup,
 /// region, deployment id, endpoints, service handles, disk registry, cancel
-/// token) into `ECStore`'s `InstanceContext`, so two `ECStore` object graphs
-/// can now stay isolated. This guard is intentionally **retained**: the startup
-/// path still publishes into the process-level *bootstrap* context (write-once
-/// region/endpoints/deployment id) and the single `GLOBAL_OBJECT_API` handle,
-/// so a second startup would fail-fast on that shared state. Lifting the guard
-/// requires threading a per-instance context through storage startup (so each
-/// server constructs its own context instead of sharing the bootstrap); until
-/// then, rejecting the second start is safer than the panic it would become.
+/// token) into `ECStore`'s `InstanceContext`, and backlog#1052 S1 threads an
+/// explicit context through the storage startup path, so a second server's
+/// *storage-layer* state could now stay isolated. This guard is still
+/// intentionally **retained**: embedded startup shares the bootstrap context
+/// (see `run_embedded_startup`), the request path resolves the store per
+/// request through the process-level `GLOBAL_OBJECT_API`/`AppContext`
+/// singletons (a second listener would serve the first instance's data), and
+/// IAM/bucket-metadata/config/credentials remain process singletons. Lifting
+/// the guard is staged in backlog#1052 (S2 per-server dispatch, S3 app
+/// subsystems, S4 node-identity globals, then S5 removes this guard); until
+/// then, rejecting the second start is safer than the failure it would become.
 static EMBEDDED_SERVER_STARTED: AtomicBool = AtomicBool::new(false);
 
 #[derive(Debug, Clone, Copy, PartialEq, Eq)]
diff --git a/rustfs/src/startup_runtime_sources.rs b/rustfs/src/startup_runtime_sources.rs
index 121ea785a..9b9a19beb 100644
--- a/rustfs/src/startup_runtime_sources.rs
+++ b/rustfs/src/startup_runtime_sources.rs
@@ -16,7 +16,7 @@ use crate::config::RustFSBufferConfig;
 use crate::runtime_sources::{
     current_outbound_tls_generation as runtime_current_outbound_tls_generation, current_replication_pool_handle,
 };
-use crate::storage_api::startup::runtime_sources::{DynReplicationPool, set_global_region, set_global_rustfs_port};
+use crate::storage_api::startup::runtime_sources::{DynReplicationPool, InstanceContext, set_global_rustfs_port};
 use rustfs_kms::KmsServiceManager;
 use rustfs_obs::{GlobalError as ObservabilityError, OtelGuard};
 use rustfs_tls_runtime::{OutboundTlsMaterial, TlsGeneration};
@@ -30,8 +30,8 @@ pub(crate) fn init_action_credentials(
     rustfs_credentials::init_global_action_credentials(Some(access_key), Some(secret_key))
 }
 
-pub(crate) fn publish_region(region: s3s::region::Region) {
-    set_global_region(region);
+pub(crate) fn publish_region(instance_ctx: &Arc<InstanceContext>, region: s3s::region::Region) {
+    instance_ctx.set_region(region);
 }
 
 pub(crate) fn publish_server_port(port: u16) {
diff --git a/rustfs/src/startup_server.rs b/rustfs/src/startup_server.rs
index 439babda8..4fe60fbac 100644
--- a/rustfs/src/startup_server.rs
+++ b/rustfs/src/startup_server.rs
@@ -17,6 +17,7 @@ use crate::{
     config::Config,
     server::{ServiceState, ServiceStateManager, ShutdownHandle, start_http_server},
     startup_runtime_sources,
+    storage_api::startup::runtime_sources::InstanceContext,
 };
 use rustfs_common::GlobalReadiness;
 use rustfs_utils::net::parse_and_resolve_address;
@@ -76,14 +77,17 @@ pub(crate) struct StartupHttpServers {
     pub(crate) console_shutdown_tx: Option<ShutdownHandle>,
 }
 
-pub(crate) async fn init_startup_listen_context(config: &Config) -> Result<StartupListenContext> {
+pub(crate) async fn init_startup_listen_context(
+    config: &Config,
+    instance_ctx: &Arc<InstanceContext>,
+) -> Result<StartupListenContext> {
     log_sanitized_server_config(config);
     let readiness = Arc::new(GlobalReadiness::new());
 
     if let Some(region_str) = &config.region {
         region_str
             .parse::<s3s::region::Region>()
-            .map(startup_runtime_sources::publish_region)
+            .map(|region| startup_runtime_sources::publish_region(instance_ctx, region))
             .map_err(|err| Error::other(format!("invalid region '{}': {}", region_str, err)))?;
     }
 
@@ -191,7 +195,10 @@ pub(crate) fn find_embedded_available_port() -> Result<u16> {
     Err(last_err.unwrap_or_else(|| Error::other("failed to reserve an embedded TCP port")))
 }
 
-pub(crate) async fn init_embedded_startup_listen_context(config: &Config) -> Result<EmbeddedStartupListenContext> {
+pub(crate) async fn init_embedded_startup_listen_context(
+    config: &Config,
+    instance_ctx: &Arc<InstanceContext>,
+) -> Result<EmbeddedStartupListenContext> {
     let readiness = Arc::new(GlobalReadiness::new());
 
     let server_addr =
@@ -210,7 +217,7 @@ pub(crate) async fn init_embedded_startup_listen_context(config: &Config) -> Res
     if let Some(region_str) = &config.region {
         region_str
             .parse::<s3s::region::Region>()
-            .map(startup_runtime_sources::publish_region)
+            .map(|region| startup_runtime_sources::publish_region(instance_ctx, region))
             .map_err(|err| Error::other(format!("invalid region '{region_str}': {err}")))?;
     }
 
diff --git a/rustfs/src/startup_storage.rs b/rustfs/src/startup_storage.rs
index 1764789e5..56e090ce4 100644
--- a/rustfs/src/startup_storage.rs
+++ b/rustfs/src/startup_storage.rs
@@ -14,9 +14,9 @@
 
 use crate::startup_fs_guard::enforce_unsupported_fs_policy;
 use crate::storage_api::startup::storage::{
-    ECStore, EndpointServerPools, global_config_init_error_is_deterministic, init_background_replication, init_ecstore_config,
-    init_global_config_sys, init_local_disks, init_lock_clients, prewarm_local_disk_id_map, set_global_endpoints,
-    try_migrate_server_config, update_erasure_type,
+    ECStore, EndpointServerPools, InstanceContext, global_config_init_error_is_deterministic, init_background_replication,
+    init_ecstore_config, init_global_config_sys, init_local_disks_with_instance_ctx, init_lock_clients,
+    prewarm_local_disk_id_map_with_instance_ctx, try_migrate_server_config,
 };
 use rustfs_common::{GlobalReadiness, SystemStage};
 use std::{
@@ -44,7 +44,11 @@ pub(crate) struct StartupStorageRuntime {
     pub(crate) shutdown_token: CancellationToken,
 }
 
-pub(crate) async fn init_startup_storage_foundation(server_address: &str, volumes: &[String]) -> Result<EndpointServerPools> {
+pub(crate) async fn init_startup_storage_foundation(
+    server_address: &str,
+    volumes: &[String],
+    instance_ctx: &Arc<InstanceContext>,
+) -> Result<EndpointServerPools> {
     info!(
         target: "rustfs::main::run",
         event = EVENT_ENDPOINT_PARSING_STARTED,
@@ -71,8 +75,8 @@ pub(crate) async fn init_startup_storage_foundation(server_address: &str, volume
         .map_err(Error::other)?;
     enforce_unsupported_fs_policy(&endpoint_pools)?;
 
-    set_global_endpoints(endpoint_pools.as_ref().clone());
-    update_erasure_type(setup_type).await;
+    instance_ctx.set_endpoints(endpoint_pools.clone());
+    instance_ctx.update_erasure_type(setup_type).await;
 
     debug!(
         target: "rustfs::main::run",
@@ -83,7 +87,7 @@ pub(crate) async fn init_startup_storage_foundation(server_address: &str, volume
         state = "starting",
         "starting local disk initialization"
     );
-    init_local_disks(endpoint_pools.clone())
+    init_local_disks_with_instance_ctx(instance_ctx, endpoint_pools.clone())
         .await
         .inspect_err(|err| {
             error!(
@@ -98,7 +102,7 @@ pub(crate) async fn init_startup_storage_foundation(server_address: &str, volume
             );
         })
         .map_err(Error::other)?;
-    prewarm_local_disk_id_map().await;
+    prewarm_local_disk_id_map_with_instance_ctx(instance_ctx).await;
     init_lock_clients(endpoint_pools.clone());
 
     log_storage_pool_layout(&endpoint_pools);
@@ -109,16 +113,17 @@ pub(crate) async fn init_startup_storage_foundation(server_address: &str, volume
 pub(crate) async fn init_embedded_startup_storage_foundation(
     server_address: &str,
     volumes: &[String],
+    instance_ctx: &Arc<InstanceContext>,
 ) -> Result<EndpointServerPools> {
     let (endpoint_pools, setup_type) = EndpointServerPools::from_volumes(server_address, volumes.to_vec())
         .await
         .map_err(|err| Error::other(format!("endpoints: {err}")))?;
     enforce_unsupported_fs_policy(&endpoint_pools).map_err(|err| Error::other(format!("unsupported fs guard: {err}")))?;
 
-    set_global_endpoints(endpoint_pools.as_ref().clone());
-    update_erasure_type(setup_type).await;
+    instance_ctx.set_endpoints(endpoint_pools.clone());
+    instance_ctx.update_erasure_type(setup_type).await;
 
-    init_local_disks(endpoint_pools.clone())
+    init_local_disks_with_instance_ctx(instance_ctx, endpoint_pools.clone())
         .await
         .map_err(|err| Error::other(format!("local disks: {err}")))?;
     init_lock_clients(endpoint_pools.clone());
@@ -130,6 +135,7 @@ pub(crate) async fn init_startup_storage_runtime(
     server_addr: SocketAddr,
     endpoint_pools: &EndpointServerPools,
     readiness: Arc<GlobalReadiness>,
+    instance_ctx: Arc<InstanceContext>,
 ) -> Result<StartupStorageRuntime> {
     let ctx = CancellationToken::new();
 
@@ -142,7 +148,7 @@ pub(crate) async fn init_startup_storage_runtime(
         state = "starting",
         "starting ECStore initialization"
     );
-    let store = ECStore::new(server_addr, endpoint_pools.clone(), ctx.clone())
+    let store = ECStore::new_with_instance_ctx(server_addr, endpoint_pools.clone(), ctx.clone(), instance_ctx)
         .await
         .inspect_err(|err| {
             error!(
@@ -172,21 +178,23 @@ pub(crate) async fn init_embedded_startup_storage_runtime(
     endpoint_pools: &EndpointServerPools,
     readiness: Arc<GlobalReadiness>,
     shutdown_token: CancellationToken,
+    instance_ctx: Arc<InstanceContext>,
 ) -> Result<StartupStorageRuntime> {
-    let store = match ECStore::new(server_addr, endpoint_pools.clone(), shutdown_token.clone()).await {
-        Ok(store) => store,
-        Err(err) => {
-            error!(
-                component = LOG_COMPONENT_EMBEDDED,
-                subsystem = LOG_SUBSYSTEM_EMBEDDED,
-                event = EVENT_EMBEDDED_STORAGE_INIT_FAILED,
-                stage = "ecstore_new",
-                error = ?err,
-                "Embedded storage initialization failed"
-            );
-            return Err(Error::other(format!("ECStore: {err}")));
-        }
-    };
+    let store =
+        match ECStore::new_with_instance_ctx(server_addr, endpoint_pools.clone(), shutdown_token.clone(), instance_ctx).await {
+            Ok(store) => store,
+            Err(err) => {
+                error!(
+                    component = LOG_COMPONENT_EMBEDDED,
+                    subsystem = LOG_SUBSYSTEM_EMBEDDED,
+                    event = EVENT_EMBEDDED_STORAGE_INIT_FAILED,
+                    stage = "ecstore_new",
+                    error = ?err,
+                    "Embedded storage initialization failed"
+                );
+                return Err(Error::other(format!("ECStore: {err}")));
+            }
+        };
 
     init_embedded_startup_storage_global_config(store.clone()).await?;
     readiness.mark_stage(SystemStage::StorageReady);
@@ -324,7 +332,9 @@ fn global_config_retry_exhausted(retry_count: usize) -> bool {
 
 #[cfg(test)]
 mod tests {
-    use super::{global_config_retry_exhausted, storage_pool_has_host_failure_risk};
+    use super::{global_config_retry_exhausted, init_embedded_startup_storage_foundation, storage_pool_has_host_failure_risk};
+    use crate::storage_api::startup::storage::{InstanceContext, bootstrap_instance_ctx};
+    use std::sync::Arc;
 
     #[test]
     fn reports_host_failure_risk_only_for_multi_drive_sets() {
@@ -333,6 +343,45 @@ mod tests {
         assert!(storage_pool_has_host_failure_risk(2));
     }
 
+    // Phase 5 follow-up (backlog#1052): the embedded storage foundation writes
+    // its topology (endpoints, erasure kind) into the explicitly passed
+    // instance context, leaving the process bootstrap context untouched — the
+    // seam a future second embedded server needs to avoid the write-once
+    // panics on shared startup state.
+    #[tokio::test]
+    async fn embedded_foundation_writes_land_on_passed_instance_ctx() {
+        let temp_dir = tempfile::tempdir().expect("create temp volume");
+        let volume = temp_dir.path().display().to_string();
+        let instance_ctx = Arc::new(InstanceContext::new());
+
+        init_embedded_startup_storage_foundation("127.0.0.1:29123", &[volume], &instance_ctx)
+            .await
+            .expect("embedded storage foundation should initialize");
+
+        assert!(
+            instance_ctx.endpoints().is_some(),
+            "the passed context must hold the parsed endpoint topology"
+        );
+        assert!(
+            instance_ctx.is_erasure_sd().await,
+            "a single local volume must record single-drive erasure on the passed context"
+        );
+
+        let bootstrap = bootstrap_instance_ctx();
+        assert!(
+            !Arc::ptr_eq(&instance_ctx, &bootstrap),
+            "the test context must be distinct from the process bootstrap context"
+        );
+        assert!(
+            bootstrap.endpoints().is_none(),
+            "the bootstrap context must not absorb topology written to an explicit context"
+        );
+        assert!(
+            !bootstrap.is_erasure_sd().await,
+            "the bootstrap context must not absorb the erasure kind written to an explicit context"
+        );
+    }
+
     #[test]
     fn global_config_retry_limit_matches_startup_policy() {
         assert!(!global_config_retry_exhausted(15));
diff --git a/rustfs/src/storage/mod.rs b/rustfs/src/storage/mod.rs
index e9edde377..b8b550949 100644
--- a/rustfs/src/storage/mod.rs
+++ b/rustfs/src/storage/mod.rs
@@ -51,28 +51,28 @@ pub(crate) use storage_api::{
     DEFAULT_READ_BUFFER_SIZE, DailyAllTierStats, DeleteOptions, DiskError, DiskInfo, DiskInfoOptions, DiskResult, DiskStore,
     DynReader, DynReplicationPool, ECStore, Endpoint, EndpointServerPools, Error, EventArgs, ExpiryState, FS, FileInfoVersions,
     FileReader, FileWriter, GetObjectReader, HashReader, LocalPeerS3Client, MetricType, NotificationSys, OBJECT_LOCK_CONFIG,
-    ObjectInfo, ObjectLockBlockReason, ObjectOptions, ObjectPartInfo, PEER_RESTSIGNAL, PEER_RESTSUB_SYS, PolicySys,
-    PoolEndpoints, PutObjReader, QuotaError, RUSTFS_META_BUCKET, RawFileInfo, ReadMultipleReq, ReadMultipleResp, ReadOptions,
-    RenameDataResp, ReplicationStats, ReplicationStatusType, Result, SERVICE_SIGNAL_REFRESH_CONFIG,
-    SERVICE_SIGNAL_RELOAD_DYNAMIC, SetupType, StorageDeletedObject, StorageDiskRpcExt, StorageError, StorageGetObjectReader,
-    StorageObjectInfo, StorageObjectOptions, StorageObjectToDelete, StoragePeerS3ClientExt, StoragePutObjReader,
-    StorageReplicationConfigExt, StorageVersioningConfigExt, TONIC_RPC_PREFIX, TierConfigMgr, UpdateMetadataOpts, VolumeInfo,
-    WalkDirOptions, WorkloadAdmissionSnapshotProviderRef, WriteEncryption, WritePlan, access_consumer, add_object_lock_years,
-    all_local_disk, all_local_disk_path, check_retention_for_modification, collect_local_metrics, compression_metadata_value,
-    contract, decode_tags, decode_tags_to_map, delete_bucket_metadata_config, disk_drive_path, disk_endpoint, ecfs_consumer,
-    ecfs_extend_consumer, ecstore_admin, ecstore_bucket, ecstore_capacity, ecstore_client, ecstore_cluster, ecstore_compression,
-    ecstore_config, ecstore_data_usage, ecstore_disk, ecstore_error, ecstore_event, ecstore_layout, ecstore_metrics,
-    ecstore_notification, ecstore_rebalance, ecstore_rio, ecstore_rpc, ecstore_set_disk, ecstore_storage, ecstore_tier,
-    encode_tags, find_local_disk_by_ref, get_bucket_accelerate_config, get_bucket_cors_config, get_bucket_logging_config,
-    get_bucket_metadata, get_bucket_notification_config, get_bucket_object_lock_config, get_bucket_policy_raw,
-    get_bucket_replication_config, get_bucket_request_payment_config, get_bucket_sse_config, get_bucket_website_config,
-    get_local_server_property, get_lock_acquire_timeout, get_public_access_block_config, head_prefix_consumer, helper_consumer,
-    init_background_replication, init_bucket_metadata_sys, init_ecstore_config, init_local_disks, init_lock_clients,
+    ObjectInfo, ObjectLockBlockReason, ObjectOptions, ObjectPartInfo, PEER_RESTSIGNAL, PEER_RESTSUB_SYS, PolicySys, PutObjReader,
+    QuotaError, RUSTFS_META_BUCKET, RawFileInfo, ReadMultipleReq, ReadMultipleResp, ReadOptions, RenameDataResp,
+    ReplicationStats, ReplicationStatusType, Result, SERVICE_SIGNAL_REFRESH_CONFIG, SERVICE_SIGNAL_RELOAD_DYNAMIC,
+    StorageDeletedObject, StorageDiskRpcExt, StorageError, StorageGetObjectReader, StorageObjectInfo, StorageObjectOptions,
+    StorageObjectToDelete, StoragePeerS3ClientExt, StoragePutObjReader, StorageReplicationConfigExt, StorageVersioningConfigExt,
+    TONIC_RPC_PREFIX, TierConfigMgr, UpdateMetadataOpts, VolumeInfo, WalkDirOptions, WorkloadAdmissionSnapshotProviderRef,
+    WriteEncryption, WritePlan, access_consumer, add_object_lock_years, all_local_disk, all_local_disk_path,
+    check_retention_for_modification, collect_local_metrics, compression_metadata_value, contract, decode_tags,
+    decode_tags_to_map, delete_bucket_metadata_config, disk_drive_path, disk_endpoint, ecfs_consumer, ecfs_extend_consumer,
+    ecstore_admin, ecstore_bucket, ecstore_capacity, ecstore_client, ecstore_cluster, ecstore_compression, ecstore_config,
+    ecstore_data_usage, ecstore_disk, ecstore_error, ecstore_event, ecstore_layout, ecstore_metrics, ecstore_notification,
+    ecstore_rebalance, ecstore_rio, ecstore_rpc, ecstore_set_disk, ecstore_storage, ecstore_tier, encode_tags,
+    find_local_disk_by_ref, get_bucket_accelerate_config, get_bucket_cors_config, get_bucket_logging_config, get_bucket_metadata,
+    get_bucket_notification_config, get_bucket_object_lock_config, get_bucket_policy_raw, get_bucket_replication_config,
+    get_bucket_request_payment_config, get_bucket_sse_config, get_bucket_website_config, get_local_server_property,
+    get_lock_acquire_timeout, get_public_access_block_config, head_prefix_consumer, helper_consumer, init_background_replication,
+    init_bucket_metadata_sys, init_ecstore_config, init_local_disks_with_instance_ctx, init_lock_clients,
     is_all_buckets_not_found, is_err_bucket_not_found, is_err_object_not_found, is_err_version_not_found, is_valid_storage_class,
-    load_bucket_metadata, options_consumer, prewarm_local_disk_id_map, read_config, record_replication_proxy, rpc_consumer,
-    runtime_sources_consumer, s3_api_consumer, save_config, serialize, set_bucket_metadata, table_catalog_path_hash, to_s3s_etag,
-    topology_snapshot_from_endpoint_pools_with_capabilities, try_migrate_bucket_metadata, try_migrate_iam_config,
-    try_migrate_server_config, update_bucket_metadata_config, verify_rpc_signature, wrap_reader,
+    load_bucket_metadata, options_consumer, prewarm_local_disk_id_map_with_instance_ctx, read_config, record_replication_proxy,
+    rpc_consumer, runtime_sources_consumer, s3_api_consumer, save_config, serialize, set_bucket_metadata,
+    table_catalog_path_hash, to_s3s_etag, topology_snapshot_from_endpoint_pools_with_capabilities, try_migrate_bucket_metadata,
+    try_migrate_iam_config, try_migrate_server_config, update_bucket_metadata_config, verify_rpc_signature, wrap_reader,
 };
 
 #[cfg(test)]
diff --git a/rustfs/src/storage/storage_api.rs b/rustfs/src/storage/storage_api.rs
index 30012fb0f..659ab6fa1 100644
--- a/rustfs/src/storage/storage_api.rs
+++ b/rustfs/src/storage/storage_api.rs
@@ -389,15 +389,14 @@ pub(crate) mod ecstore_event {
 
 pub(crate) mod ecstore_global {
     pub(crate) use rustfs_ecstore::api::global::{
-        set_global_endpoints, set_global_region, set_global_rustfs_port, set_object_store_resolver, shutdown_background_services,
-        update_erasure_type,
+        set_global_rustfs_port, set_object_store_resolver, shutdown_background_services,
     };
 }
 
 pub(crate) mod ecstore_runtime {
     pub(crate) use rustfs_ecstore::api::runtime::{
-        boot_time, bucket_monitor, deployment_id, endpoint_pools, global_lock_client, global_lock_clients,
-        global_tier_config_mgr, region, rustfs_port, setup_is_dist_erasure,
+        InstanceContext, boot_time, bootstrap_ctx, bucket_monitor, deployment_id, endpoint_pools, global_lock_client,
+        global_lock_clients, global_tier_config_mgr, region, rustfs_port, setup_is_dist_erasure,
     };
 }
 
@@ -450,9 +449,11 @@ pub(crate) mod ecstore_set_disk {
 }
 
 pub(crate) mod ecstore_storage {
+    #[cfg(test)]
+    pub(crate) use rustfs_ecstore::api::storage::init_local_disks;
     pub(crate) use rustfs_ecstore::api::storage::{
-        ECStore, all_local_disk, all_local_disk_path, find_local_disk_by_ref, init_local_disks, init_lock_clients,
-        prewarm_local_disk_id_map,
+        ECStore, all_local_disk, all_local_disk_path, find_local_disk_by_ref, init_local_disks_with_instance_ctx,
+        init_lock_clients, prewarm_local_disk_id_map_with_instance_ctx,
     };
 }
 
@@ -516,13 +517,15 @@ pub(crate) type FileReader = ecstore_disk::FileReader;
 pub(crate) type FileWriter = ecstore_disk::FileWriter;
 pub(crate) type FS = super::ecfs::FS;
 pub(crate) type HashReader = ecstore_rio::HashReader;
+pub(crate) type InstanceContext = ecstore_runtime::InstanceContext;
 pub(crate) type LocalPeerS3Client = ecstore_rpc::LocalPeerS3Client;
 pub(crate) type MetricType = ecstore_metrics::MetricType;
 pub(crate) type ObjectPartInfo = rustfs_filemeta::ObjectPartInfo;
 pub(crate) type ObjectLockBlockReason = ecstore_bucket::object_lock::objectlock_sys::ObjectLockBlockReason;
 pub(crate) type ObjectStoreResolver = dyn Fn() -> Option<Arc<ECStore>> + Send + Sync + 'static;
-pub(crate) type PolicySys = ecstore_bucket::policy_sys::PolicySys;
+#[cfg(test)]
 pub(crate) type PoolEndpoints = ecstore_layout::PoolEndpoints;
+pub(crate) type PolicySys = ecstore_bucket::policy_sys::PolicySys;
 pub(crate) type WorkloadAdmissionSnapshotProviderRef = rustfs_ecstore::WorkloadAdmissionSnapshotProviderRef;
 pub(crate) type QuotaError = ecstore_bucket::quota::QuotaError;
 pub(crate) type RawFileInfo = rustfs_filemeta::RawFileInfo;
@@ -535,7 +538,6 @@ pub(crate) type OldCurrentSize = ecstore_disk::OldCurrentSize;
 pub(crate) type RenameDataResp = ecstore_disk::RenameDataResp;
 pub(crate) type ReplicationStatusType = ecstore_bucket::replication::ReplicationStatusType;
 pub(crate) type ReplicationStats = StorageReplicationStatsHandle;
-pub(crate) type SetupType = ecstore_layout::SetupType;
 pub(crate) type StorageError = ecstore_error::StorageError;
 pub(crate) type TierConfigMgr = ecstore_tier::TierConfigMgr;
 pub(crate) use ecstore_disk::validate_batch_read_version_item_count;
@@ -757,10 +759,26 @@ pub(crate) fn global_config_init_error_is_deterministic(err: &Error) -> bool {
     ecstore_config::com::is_server_config_corrupt_error(err)
 }
 
+pub(crate) async fn init_local_disks_with_instance_ctx(
+    instance_ctx: &Arc<InstanceContext>,
+    endpoint_pools: EndpointServerPools,
+) -> Result<()> {
+    ecstore_storage::init_local_disks_with_instance_ctx(instance_ctx, endpoint_pools).await
+}
+
+/// Test-only legacy entry: registers disks into the ambient (process-default)
+/// context, matching what pre-#1052 production startup did.
+#[cfg(test)]
 pub(crate) async fn init_local_disks(endpoint_pools: EndpointServerPools) -> Result<()> {
     ecstore_storage::init_local_disks(endpoint_pools).await
 }
 
+/// The process-level bootstrap instance context that single-instance startup
+/// threads through the storage foundation (Phase 5 follow-up, backlog#1052).
+pub(crate) fn bootstrap_instance_ctx() -> Arc<InstanceContext> {
+    ecstore_runtime::bootstrap_ctx()
+}
+
 pub(crate) fn init_lock_clients(endpoint_pools: EndpointServerPools) {
     ecstore_storage::init_lock_clients(endpoint_pools);
 }
@@ -773,8 +791,8 @@ pub(crate) async fn read_config(api: Arc<ECStore>, file: &str) -> Result<Vec<u8>
     ecstore_config::com::read_config(api, file).await
 }
 
-pub(crate) async fn prewarm_local_disk_id_map() {
-    ecstore_storage::prewarm_local_disk_id_map().await;
+pub(crate) async fn prewarm_local_disk_id_map_with_instance_ctx(instance_ctx: &Arc<InstanceContext>) {
+    ecstore_storage::prewarm_local_disk_id_map_with_instance_ctx(instance_ctx).await;
 }
 
 pub(crate) fn replication_queue_current_count() -> Option<i64> {
@@ -793,14 +811,6 @@ pub(crate) fn shutdown_background_monitors() {
     rustfs_ecstore::shutdown_background_monitors();
 }
 
-pub(crate) fn set_global_endpoints(endpoints: Vec<PoolEndpoints>) {
-    ecstore_global::set_global_endpoints(endpoints);
-}
-
-pub(crate) fn set_global_region(region: s3s::region::Region) {
-    ecstore_global::set_global_region(region);
-}
-
 pub(crate) fn set_global_rustfs_port(value: u16) {
     ecstore_global::set_global_rustfs_port(value);
 }
@@ -815,10 +825,6 @@ pub(crate) async fn try_migrate_server_config(store: Arc<ECStore>) {
     ecstore_config::try_migrate_server_config(store, Some(decrypt_fn)).await;
 }
 
-pub(crate) async fn update_erasure_type(setup_type: SetupType) {
-    ecstore_global::update_erasure_type(setup_type).await;
-}
-
 pub(crate) trait StorageDiskRpcExt {
     async fn disk_info(&self, opts: &DiskInfoOptions) -> DiskResult<DiskInfo>;
     async fn delete_volume(&self, volume: &str, force_delete: bool) -> DiskResult<()>;
diff --git a/rustfs/src/storage_api.rs b/rustfs/src/storage_api.rs
index 4f83c6611..19bb24bc2 100644
--- a/rustfs/src/storage_api.rs
+++ b/rustfs/src/storage_api.rs
@@ -211,7 +211,7 @@ pub(crate) mod startup {
     }
 
     pub(crate) mod runtime_sources {
-        pub(crate) use crate::storage::storage_api::{DynReplicationPool, set_global_region, set_global_rustfs_port};
+        pub(crate) use crate::storage::storage_api::{DynReplicationPool, InstanceContext, set_global_rustfs_port};
     }
 
     pub(crate) mod services {
@@ -226,9 +226,10 @@ pub(crate) mod startup {
 
     pub(crate) mod storage {
         pub(crate) use crate::storage::storage_api::{
-            ECStore, EndpointServerPools, global_config_init_error_is_deterministic, init_background_replication,
-            init_compression_total_memory_from_backend, init_ecstore_config, init_global_config_sys, init_local_disks,
-            init_lock_clients, prewarm_local_disk_id_map, set_global_endpoints, try_migrate_server_config, update_erasure_type,
+            ECStore, EndpointServerPools, InstanceContext, bootstrap_instance_ctx, global_config_init_error_is_deterministic,
+            init_background_replication, init_compression_total_memory_from_backend, init_ecstore_config, init_global_config_sys,
+            init_local_disks_with_instance_ctx, init_lock_clients, prewarm_local_disk_id_map_with_instance_ctx,
+            try_migrate_server_config,
         };
     }
 }