disksToHeal = append(disksToHeal, disk.Endpoint())
		}
	}
	if len(disksToHeal) == globalEndpoints.NEndpoints() {
		// When all disks == all command line endpoints
		// this is a fresh setup, no need to trigger healing.
		return Endpoints{}
	}
	return disksToHeal
}

var newDiskHealingTimeout = newDynamicTimeout(30*time.Second, 10*time.Second)

```
sudo pip install yq
alias kv-pairify='yq ".configuration[]" | jq ".[]" | jq -r ".name + \"=\" + .value"'
```

Let's take for example a set of 12 compute nodes with an aggregate memory of _1.2TiB_, we need to do following settings for optimal results. Add the following optimal entries for _core-site.xml_ to configure _s3a_ with **MinIO**. Most important options here are

```

  private final String subtype;
  private final ImmutableListMultimap<String, String> parameters;

  @LazyInit private @Nullable String toString;

  @LazyInit private int hashCode;

  // We need to differentiate between "not computed" and "computed to be absent."
  @SuppressWarnings("NullableOptional")
  @LazyInit
  private @Nullable Optional<Charset> parsedCharset;

        for (SmbTransportImpl transport : connections) {
            try {
                // Check if connection has been idle too long
                // Note: getLastUseTime() method would need to be added to SmbTransportImpl
                // For now, we'll use a simple disconnected check
                if (transport.isDisconnected()) {
                    idle.add(transport);

 *       hunt through classes like {@link Arrays} and {@link Longs} for them.
 *   <li>Supports a copy-free {@link #subArray} view, so methods that accept this type don't need to
 *       add overloads that accept start and end indexes.
 *   <li>Can be streamed without "breaking the chain": {@code foo.getBarLongs().stream()...}.

   * matches.
   *
   * <p>On the other hand, if for example we are resolving {@code List<A extends B>} to {@code
   * List<A extends String>}, we need to compare that {@code <A extends B>} is unequal to {@code <A
   * extends String>} in order to decide to use the transformed type instead of the original type.
   */
  static final class TypeVariableKey {

 *       hunt through classes like {@link Arrays} and {@link Ints} for them.
 *   <li>Supports a copy-free {@link #subArray} view, so methods that accept this type don't need to
 *       add overloads that accept start and end indexes.
 *   <li>Can be streamed without "breaking the chain": {@code foo.getBarInts().stream()...}.

      @SuppressWarnings({"rawtypes", "unchecked"})
      Supplier<List<V>> result = (Supplier) INSTANCE;
      return result;
    }

    @Override
    // We recommend against linkedListValues but need to keep it for compatibility.
    @SuppressWarnings("JdkObsolete")
    public List<?> get() {
      return new LinkedList<>();
    }
  }

  private static final class HashSetSupplier<V extends @Nullable Object>

   * is {@code 60}, and the concurrency level is {@code 8}, then eight segments are created, each
   * having a hash table of size eight. Providing a large enough estimate at construction time
   * avoids the need for expensive resizing operations later, but setting this value unnecessarily
   * high wastes memory.
   *
   * @return this {@code CacheBuilder} instance (for chaining)

      }

      static class GistFile {
        String content;
      }
    ```

### Response Caching ([.kt][CacheResponseKotlin], [.java][CacheResponseJava])

To cache responses, you'll need a cache directory that you can read and write to, and a limit on the cache's size. The cache directory should be private, and untrusted applications should not be able to read its contents!