For large responses, returning a `Response` directly is much faster than returning a dictionary.

This is because by default, FastAPI will inspect every item inside and make sure it is serializable as JSON, using the same [JSON Compatible Encoder](../tutorial/encoder.md){.internal-link target=_blank} explained in the tutorial. This is what allows you to return **arbitrary objects**, for example database models.

## The `password` flow

Now let's go back a bit and understand what is all that.

The `password` "flow" is one of the ways ("flows") defined in OAuth2, to handle security and authentication.

OAuth2 was designed so that the backend or API could be independent of the server that authenticates the user.

(A SQL separator is semicolon ";")

take-finally-*.sql:
SQL statements for check loaded data (or DDL after data loading)
You should write your own SQL statements in this file.
(basically same specifications as replace-schema.sql)

The "data" directory is for data loading like this:
/- - - - - - - - - - - - - - - - - - - -
playsql
  |-data
     |-common
     |  |-xls
     |     |-10-master.xls

  private lateinit var server: MockWebServer

  @BeforeEach
  fun setUp(server: MockWebServer) {
    this.server = server

    // Default after JDK 14, but we are avoiding tests that assume special setup.
    // System.setProperty("jdk.tls.client.enableSessionTicketExtension", "true")
    // System.setProperty("jdk.tls.server.enableSessionTicketExtension", "true")

    platform.assumeJdk9()
  }

  @Test

    return Collections.unmodifiableSet(outEdgeMap.keySet());
  }

  @Override
  public N adjacentNode(E edge) {
    // Since the reference node is defined to be 'source' for directed graphs,
    // we can assume this edge lives in the set of outgoing edges.
    // (We're relying on callers to call this method only with an edge that's in the graph.)
    return requireNonNull(outEdgeMap.get(edge));
  }

  @Override

`kube.Client` just bundles these all in one place, so we have a single object to pass around anywhere we need Kubernetes access.
Additionally, it has a fake client variant, that does the same thing but with fake clients for use in unit tests.

Aside from this, there are a few conveniences and workarounds built-in to the client to make things simpler.

64 sets, which is *16* 64 = 1024* drives in total.

- *If total number of nodes are of odd number then GCD algorithm provides affinity towards odd number erasure sets to provide for uniform distribution across nodes*. This is to ensure that same number of drives are pariticipating in any erasure set. For example if you have 2 nodes with 180 drives then GCD is 15 but this would lead to uneven distribution, one of the nodes would participate more drives. To avoid this the affinity is given towards...

  // name represents the name of the CSI driver that this object refers to.
  // This MUST be the same name returned by the CSI GetPluginName() call for
  // that driver.
  optional string name = 1;

  // nodeID of the node from the driver point of view.
  // This field enables Kubernetes to communicate with storage systems that do
  // not share the same nomenclature for nodes. For example, Kubernetes may

                should have all the fields, including the ones that have the same value
                as the default. This is different from `response_model_exclude_unset`
                in that if the fields are set but contain the same default values,
                they will be excluded from the response.

 *       {@link Hasher}), but this is merely a convenience; these are always translated into raw
 *       byte sequences under the covers.
 *   <li><b>hash code:</b> each hash function always yields hash codes of the same fixed bit length
 *       (given by {@link #bits}). For example, {@link Hashing#sha1} produces a 160-bit number,
 *       while {@link Hashing#murmur3_32()} yields only 32 bits. Because a {@code long} value is