### Consistency Guarantees

MinIO follows strict **read-after-write** and **list-after-write** consistency model for all i/o operations both in distributed and standalone modes. This consistency model is only guaranteed if you use disk filesystems such as xfs, zfs or btrfs etc.. for distributed setup.

### Creating the `Settings` only once with `lru_cache` { #creating-the-settings-only-once-with-lru-cache }

Reading a file from disk is normally a costly (slow) operation, so you probably want to do it only once and then reuse the same settings object, instead of reading it for each request.

But every time we do:

```Python
Settings()
```

* Adding vmdk file extension for vmDiskPath in vsphere DeleteVolume ([#40538](https://github.com/kubernetes/kubernetes/pull/40538), [@divyenpatel](https://github.com/divyenpatel))
* Prevent hotloops on error conditions, which could fill up the disk faster than log rotation can free space. ([#40497](https://github.com/kubernetes/kubernetes/pull/40497), [@lavalamp](https://github.com/lavalamp))

	}
	return clnt
}

// IsOnline returns whether the client is likely to be online.
func (c *Client) IsOnline() bool {
	return atomic.LoadInt32(&c.connected) == online
}

// LastConn returns when the disk was (re-)connected
func (c *Client) LastConn() time.Time {
	return time.Unix(0, atomic.LoadInt64(&c.lastConn))
}

// LastError returns previous error
func (c *Client) LastError() error {
	c.RLock()

                    "DHnC", // Durable handle reconnect
                    "AlSi", // Allocation size
                    "MxAc", // Max access
                    "TWrp", // Timewarp
                    "QFid", // Query on disk ID
                    "RqLs", // Request lease
                    "DH2Q", // Durable handle request V2
                    "DH2C", // Durable handle reconnect V2
                    "ExtA", // Extended attributes

        // The root artifact may, or may not be resolved so we need to check before we attempt to resolve.
        // This is often an artifact like a POM that is taken from disk and we already have hold of the
        // file reference. But this may be a Maven Plugin that we need to resolve from a remote repository
        // as well as its dependencies.

	}

	sort.Slice(result, func(i, j int) bool {
		return result[i].Name < result[j].Name
	})

	return result, nil
}

// GetBucketInfo returns bucket stat info about bucket on disk across all peers
func (sys *S3PeerSys) GetBucketInfo(ctx context.Context, bucket string, opts BucketOptions) (binfo BucketInfo, err error) {
	g := errgroup.WithNErrs(len(sys.peerClients))

	rx.sma.addSample(rx.Curr)
	rx.Avg = rx.sma.simpleMovingAvg()
	if rx.Curr > rx.Peak {
		rx.Peak = rx.Curr
	}
	rx.N++
}

// ReplicationMRFStats holds stats of MRF backlog saved to disk in the last 5 minutes
// and number of entries that failed replication after 3 retries
type ReplicationMRFStats struct {
	LastFailedCount uint64 `json:"failedCount_last5min"`

- Improved the `FreeDiskSpaceFailed` warning event to provide more actionable details when image garbage collection fails to free enough disk space. Example: `Insufficient free disk space on the node's image filesystem (95.0% of 10.0 GiB used). Failed to free sufficient space by deleting unused images. Consider resizing the disk or deleting unused files.`. ([#132578](https://github.com/kubernetes/kubernetes/pull/132578), [@drigz](https://github.com/drigz))

	// to the in-memory cache since the disk loading began. If there
	// were changes to the in-memory cache we should wait for the next
	// cycle until we can safely update the in-memory cache.
	//
	// An in-memory cache must be replaced only if we know for sure that the
	// values loaded from disk are not stale. They might be stale if the