fix: deflake //rs/tests/nested/nns_recovery:nr_local (#9303)

## Root Cause

The test `//rs/tests/nested/nns_recovery:nr_local` flaked with:

```
Could not create message canister.: "Timeout while creating message canister"
```

`store_message` creates the canister with a single attempt that can time
out under load.

## Fix

Replace `store_message` with `store_message_with_retries`, which retries
canister creation for up to 300s with 10s intervals, making it resilient
to transient timeouts.

---

This PR was created following the steps in
`.claude/skills/fix-flaky-tests/SKILL.md`.

Author: basvandijk

rs/tests/nested/nns_recovery/common.rs

@@ -9,7 +9,7 @@ use ic_consensus_system_test_subnet_recovery::utils::{
 use ic_consensus_system_test_utils::{
     impersonate_upstreams,
     node::await_subnet_earliest_topology_version_with_retries,
-    rw_message::store_message,
+    rw_message::store_message_with_retries,
     ssh_access::{
         AuthMean, disable_ssh_access_to_node, get_updatesubnetpayload_with_keys,
         update_subnet_record, wait_until_authentication_is_granted,
@@ -215,7 +215,7 @@ pub fn test(env: TestEnv, cfg: TestConfig) {
 
     info!(logger, "Ensure NNS subnet is functional");
     let init_msg = "subnet recovery works!";
-    let app_can_id = store_message(
+    let app_can_id = store_message_with_retries(
         &nns_node.get_public_url(),
         nns_node.effective_canister_id(),
         init_msg,