Merge #66041

66041: roachtest: increase restore concurrency in restore2TB/nodes=6 r=pbardea a=pbardea

This roachtest runs on a smaller cluster with larger nodes (16 vCPU and
larger disks). This change increases the concurrency this test since
these nodes are expected to have more memory and thus can comfortably
sustain more parallel workers.

Release note: None

Co-authored-by: Paul Bardea <[email protected]>

pkg/cmd/roachtest/restore.go

@@ -408,9 +408,20 @@ func registerRestore(r *testRegistry) {
 					defer dul.Done()
 					defer hc.Done()
 					t.Status(`running restore`)
-					// Tick once before starting the restore, and once after to capture the
-					// total elapsed time. This is used by roachperf to compute and display
-					// the average MB/sec per node.
+					// Tick once before starting the restore, and once after to
+					// capture the total elapsed time. This is used by
+					// roachperf to compute and display the average MB/sec per
+					// node.
+					if item.cpus >= 16 {
+						// If the nodes are large enough (specifically, if they
+						// have enough memory we can increase the parallelism
+						// of restore). Machines with 16 vCPUs typically have
+						// enough memory to supoprt 3 concurrent workers.
+						c.Run(ctx, c.Node(1),
+							`./cockroach sql --insecure -e "SET CLUSTER SETTING kv.bulk_io_write.restore_node_concurrency = 3"`)
+						c.Run(ctx, c.Node(1),
+							`./cockroach sql --insecure -e "SET CLUSTER SETTING kv.bulk_io_write.concurrent_addsstable_requests = 3"`)
+					}
 					tick()
 					item.dataSet.runRestore(ctx, c)
 					tick()