* @author ******@****.*** (Geoff Pike)
 */
final class Fingerprint2011 extends AbstractNonStreamingHashFunction {
  static final HashFunction FINGERPRINT_2011 = new Fingerprint2011();

  // Some primes between 2^63 and 2^64 for various uses.
  private static final long K0 = 0xa5b85c5e198ed849L;
  private static final long K1 = 0x8d58ac26afe12e47L;
  private static final long K2 = 0xc47b6e9e3a970ed3L;

 * @author ******@****.*** (Geoff Pike)
 */
final class Fingerprint2011 extends AbstractNonStreamingHashFunction {
  static final HashFunction FINGERPRINT_2011 = new Fingerprint2011();

  // Some primes between 2^63 and 2^64 for various uses.
  private static final long K0 = 0xa5b85c5e198ed849L;
  private static final long K1 = 0x8d58ac26afe12e47L;
  private static final long K2 = 0xc47b6e9e3a970ed3L;

# Erasure code sizing guide

## Toy Setups

Capacity constrained environments, MinIO will work but not recommended for production.

| servers | drives (per node) | stripe_size | parity chosen (default) | tolerance for reads (servers) | tolerance for writes (servers) |
|--------:|------------------:|------------:|------------------------:|------------------------------:|-------------------------------:|

	ch := make(chan madmin.SpeedTestResult, 1)
	go func() {
		defer xioutil.SafeClose(ch)

		concurrency := opts.concurrencyStart

		if opts.autotune {
			// if we have less drives than concurrency then choose
			// only the concurrency to be number of drives to start
			// with - since default '32' might be big and may not
			// complete in total time of 10s.
			if globalEndpoints.NEndpoints() < concurrency {

Framing rules make it impractical to implement http/2 correctly on a single blocking thread. The flow-control features introduce feedback between reads and writes, requiring writes to acknowledge reads and reads to throttle writes.

In OkHttp we expose a blocking API over a framed protocol. This document explains the code and policy that makes that work.

### Threads

#### Application's calling thread

	driveOfflineCountMD = NewGaugeMD(driveOfflineCount,
		"Count of offline drives")
	driveOnlineCountMD = NewGaugeMD(driveOnlineCount,
		"Count of online drives")
	driveCountMD = NewGaugeMD(driveCount,
		"Count of all drives")

	// iostat related
	driveReadsPerSecMD = NewGaugeMD(driveReadsPerSec,
		"Reads per second on a drive",
		allDriveLabels...)

describes everything that can be configured in a MinIO setup, such as '--address', '--console-address' and command line arguments for the MinIO server.

Historically everything to MinIO was provided via command arguments for the hostnames and the drives via an ellipses syntax such as `minio server http://host{1...4}/disk{1...4}` this requirement added an additional burden to have sequential hostnames for us to make sure that we can provide horizontal distribution, however we have come across situations...

- We limited the number of drives to 16 for erasure set because, erasure code shards more than 16 can become chatty and do not have any performance advantages. Additionally since 16 drive erasure set gives you tolerance of 8 drives per object by default which is plenty in any practical scenario.

on 16 drive MinIO deployment. If you use eight data and eight parity drives, the file space usage will be approximately twice, i.e. 100 MiB
file will take 200 MiB space. But, if you use ten data and six parity drives, same 100 MiB file takes around 160 MiB. If you use 14 data and
two parity drives, 100 MiB file takes only approximately 114 MiB.

A stand-alone MinIO server would go down if the server hosting the drives goes offline. In contrast, a distributed MinIO setup with _m_ servers and _n_ drives will have your data safe as long as _m/2_ servers or _m*n_/2 or more drives are online.

For example, an 16-server distributed setup with 200 drives per node would continue serving files, up to 4 servers can be offline in default configuration i.e around 800 drives down MinIO would continue to read and write objects.