## Upgrading MinIO

Upgrades require zero downtime in MinIO, all upgrades are non-disruptive, all transactions on MinIO are atomic. So upgrading all the servers simultaneously is the recommended way to upgrade MinIO.

	if err != nil {
		err = msgp.WrapError(err, "N")
		return
	}
	return
}

// MarshalMsg implements msgp.Marshaler
func (z AccElem) MarshalMsg(b []byte) (o []byte, err error) {
	o = msgp.Require(b, z.Msgsize())
	// map header, size 3
	// string "Total"
	o = append(o, 0x83, 0xa5, 0x54, 0x6f, 0x74, 0x61, 0x6c)
	o = msgp.AppendInt64(o, z.Total)
	// string "Size"
	o = append(o, 0xa4, 0x53, 0x69, 0x7a, 0x65)

			err = msgp.WrapError(err, "Rules", za0001)
			return
		}
	}
	return
}

// MarshalMsg implements msgp.Marshaler
func (z *BatchJobExpire) MarshalMsg(b []byte) (o []byte, err error) {
	o = msgp.Require(b, z.Msgsize())
	// map header, size 6
	// string "APIVersion"
	o = append(o, 0x86, 0xaa, 0x41, 0x50, 0x49, 0x56, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e)
	o = msgp.AppendString(o, z.APIVersion)
	// string "Bucket"

workloads typically interact with object stores in the same way they do with HDFS. These workloads rely on HDFS atomic rename functionality to complete writing data to the datastore. Object storage operations are atomic by nature and they do not require/implement rename API. The default S3A committer emulates renames through copy and delete APIs. This interaction pattern causes significant loss of performance because of the write amplification. _Netflix_, for example, developed two new staging committers...

			err = msgp.WrapError(err, "Expire")
			return
		}
	}
	return
}

// MarshalMsg implements msgp.Marshaler
func (z *BatchJobRequest) MarshalMsg(b []byte) (o []byte, err error) {
	o = msgp.Require(b, z.Msgsize())
	// map header, size 6
	// string "ID"
	o = append(o, 0x86, 0xa2, 0x49, 0x44)
	o = msgp.AppendString(o, z.ID)
	// string "User"
	o = append(o, 0xa4, 0x55, 0x73, 0x65, 0x72)

}

// AbortMultipartUpload - aborts an ongoing multipart operation
// signified by the input uploadID. This is an atomic operation
// doesn't require clients to initiate multiple such requests.
//
// All parts are purged from all disks and reference to the uploadID
// would be removed from the system, rollback is not possible on this
// operation.

			err = msgp.WrapError(err, zb0004)
			return
		}
	}
	return
}

// MarshalMsg implements msgp.Marshaler
func (z dataUsageHashMap) MarshalMsg(b []byte) (o []byte, err error) {
	o = msgp.Require(b, z.Msgsize())
	o = msgp.AppendArrayHeader(o, uint32(len(z)))
	for zb0004 := range z {
		o = msgp.AppendString(o, zb0004)
	}
	return
}

// UnmarshalMsg implements msgp.Unmarshaler

cryptographically secure pseudo-random number generator ([CSPRNG](#csprng)). However, in the context of SSE, the MinIO server does not require that the CSPRNG generates values that are indistinguishable from truly random bit strings. Instead, it is sufficient if the generated values are unique - which is a weaker requirement. Nevertheless other parts - for example the TLS-stack - may require that CSPRNG-generated values are indistinguishable from truly random bit strings.

### Server-Side Encryption with...

				err = msgp.WrapError(err, zb0006, zb0008)
				return
			}
		}
	}
	return
}

// MarshalMsg implements msgp.Marshaler
func (z localLockMap) MarshalMsg(b []byte) (o []byte, err error) {
	o = msgp.Require(b, z.Msgsize())
	o = msgp.AppendMapHeader(o, uint32(len(z)))
	for zb0006, zb0007 := range z {
		o = msgp.AppendString(o, zb0006)
		o = msgp.AppendArrayHeader(o, uint32(len(zb0007)))
		for zb0008 := range zb0007 {

          reasonable to satisfy the conditions by providing a URI or
          hyperlink to a resource that includes the required
          information.

       3. If requested by the Licensor, You must remove any of the
          information required by Section 3(a)(1)(A) to the extent
          reasonably practicable.

       4. If You Share Adapted Material You produce, the Adapter's