// Offset should always be aligned to b.shardSize
		// Can never happen unless there are programmer bugs
		return 0, errUnexpected
	}
	if b.rc == nil {
		// For the first ReadAt() call we need to open the stream for reading.
		b.currOffset = offset
		streamOffset := (offset/b.shardSize)*int64(b.h.Size()) + offset
		if len(b.data) == 0 && b.tillOffset != streamOffset {

## Test the setup

The next time the backup job runs, you can use the  `mc admin trace myminio` command and verify traffic is flowing to the MinIO nodes. For Veeam Backup and Replication you will need to wait for the backup to complete to the performance tier before it migrates data to the capacity tier (i.e., MinIO).

```

- To be able to easily get the temporary credentials to upload to a prefix. Make it possible for a client to upload a whole folder using the session. The server side applications need not create a presigned URL and serve to the client for each file. Since, the client would have the session it can do it by itself.

To transition objects in a bucket to a destination bucket on a different cluster, applications need to specify a transition tier defined on MinIO instead of storage class while setting up the ILM lifecycle rule.

> To create a transition tier for transitioning objects to a prefix `testprefix` in `azurebucket` on Azure blob using `mc`:

	for _, collectorPath := range h.metricsData.collectorPaths {
		if collectorPath.isDescendantOf(path) {
			gatherer := h.metricsData.mgGatherers[collectorPath]

			// For Bucket metrics we need to set the buckets argument inside the
			// metric group, so that it will affect collection. If no buckets
			// are provided, we will not return bucket metrics.
			if bmg, ok := h.metricsData.bucketMGMap[collectorPath]; ok {

            <multiplicity>*</multiplicity>
          </association>
          <comment>
            This may be removed or relocated in the near
            future. It is undecided whether plugins really need a remote
            repository set of their own.
          </comment>
        </field>
        <field xdoc.separator="blank">
          <name>profiles</name>
          <version>1.0.0+</version>

	for _, v := range ahs.healStatus {
		dst[v.Endpoint] = v.toHealingDisk()
	}

	return dst
}

// getHealLocalDiskEndpoints() returns the list of disks that need
// to be healed but there is no healing routine in progress on them.
func (ahs *allHealState) getHealLocalDiskEndpoints() Endpoints {
	ahs.RLock()
	defer ahs.RUnlock()

	var endpoints Endpoints

    <!--
      | We may potentially want to export these, but right now I'm not sure that anything Guice specific needs
      | to be made available to plugin authors. If we find people are getting fancy and want to take advantage of
      | Guice specifics we can expose that later. Really some testing needs to be done to see full hiding
      | of Guice has any impact on what we may categorize as a standard JSR-330 based Tesla/Maven plugin.

	// HTTP request for testing when `objectLayer` is set to `nil`.
	// There is no need to use an existing bucket and valid input for creating the request
	// since the `objectLayer==nil`  check is performed before any other checks inside the handlers.

From the list of deployment concepts from above, using workers would mainly help with the **replication** part, and a little bit with the **restarts**, but you still need to take care of the others:

* **Security - HTTPS**
* **Running on startup**
* ***Restarts***
* Replication (the number of processes running)
* **Memory**
* **Previous steps before starting**