test/integration/mesh.test.ts

@@ -124,7 +124,17 @@ async function settled(node: string, withinMs = 240_000): Promise<void> {
   const until = Date.now() + withinMs;
   let last = "";
   while (Date.now() < until) {
-    const said = await mesh("status --json");
+    // **Could not ask** and **asked, and the answer was bad** are different facts, and only the
+    // second is this machine's fault. The control plane is a container on the node being polled:
+    // while it applies a declaration, an exec into it can lose its fifo to containerd, and a poll
+    // loop that treats that as a verdict reports the mesh broken because the question missed.
+    const { out: said, ok } = await on("anchor",
+      `docker exec mesh-control /mesh-control status --json`);
+    if (!ok) {
+      last = said;
+      await new Promise((r) => setTimeout(r, 5000));
+      continue;
+    }
     const state = JSON.parse(said) as {
       wrong: { node: string; outcome: string; refused?: string;
                failed?: { id: string; error: string }[] }[];
@@ -138,9 +148,13 @@ async function settled(node: string, withinMs = 240_000): Promise<void> {
     }
     if (!state.waiting.some((w) => w.node === node)) return;
     last = said;
-    await new Promise((r) => setTimeout(r, 2000));
+    // Unhurried on purpose: each poll is an exec into a container on the machine that is busy
+    // applying, and asking four times a minute rather than thirty is the difference between
+    // observing the apply and competing with it.
+    await new Promise((r) => setTimeout(r, 5000));
   }
-  throw new Error(`${node} never caught up with what it was sent:\n${last}`);
+  throw new Error(`${node} never caught up with what it was sent within ` +
+    `${Math.round(withinMs / 1000)}s. Last answer, or the reason there was none:\n${last}`);
 }
 
 /**