> For the complete documentation index, see [llms.txt](https://docs.weka.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.weka.io/4.4/operation-guide/background-tasks.md).
# Background tasks
Learn how the WEKA system runs asynchronous and maintenance operations in the background, what limits apply to them, and how to monitor and control them.
Background tasks handle work that must not interfere with serving IO. Examples include checking metadata integrity, uploading and downloading snapshots, detaching an object store, reducing data, and applying a directory quota to existing data.
## Resource consumption
The WEKA system limits background tasks to 5% of the overall CPU. When the CPU is idle, background tasks can use more than the configured resources, and release them immediately when needed to serve IO.
## Concurrency limits
Two separate limits apply to background tasks:
**Queued tasks:** the number of tasks that can exist at the same time. A paused or aborted task still occupies a slot. When the queue is full, a new operation fails rather than waits.
**Running tasks:** the number of queued tasks that perform work at the same time. Additional tasks remain queued until a slot frees.
Both limits apply cluster-wide.
Task group
Maximum queued
Maximum running
Filesystem tasks
16
Set by the task-specific rules below
Directory quota tasks
32
4
If an operation cannot start because the queue is full, it fails with the following message:
```
Operation cannot start because there are already 32 tasks running
```
Wait for a running task to finish, then retry.
### Task-specific rules
**Snapshot upload and download:**
* Only a single local upload can exist concurrently inside a filesystem.
* Only a single remote upload inside a filesystem can be done concurrently. Local and remote uploads can co-exist.
* Only a single upload from any filesystem can exist in the same object store bucket, to prevent uploads from slowing each other down.
* An object store snapshot download operation cannot run at the same time as another snapshot download or upload operation.
**Data reduction:**
* Only one data reduction or data defragmentation task runs in the cluster at a time. If either task exists, no new task of either type starts.
* The cluster starts these tasks automatically. You do not create them.
**Snapshot metadata prefetch:**
* When a snapshot is downloaded from the object store, the system automatically prefetches its metadata as the initial step.
{% hint style="info" %}
Further restrictions exist between different tasks and between multiple tasks of the same type. When a background task does not run because of a restriction, the system provides a relevant message.
{% endhint %}
## Background tasks list
| Task name | Description | Possible actions |
| -------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------------- |
| OBS\_DETACH2 | Detaches an object store from a filesystem. | Pause, Resume, Abort until the task reaches the CLEAR\_MANIFEST phase |
| STOW\_UPLOAD | Uploads a snapshot from a filesystem to an object store bucket. | Pause, Resume, Abort |
| STOW\_DOWNLOAD\_FILESYSTEM | Downloads a filesystem from a locator in an object store. | Pause, Resume |
| STOW\_DOWNLOAD\_SNAPSHOT | Downloads a snapshot to a filesystem from a locator in an object store. Includes fetching the snapshot metadata and squashing the filesystem. | Pause, Resume |
| SNAPSHOT\_DELETE | Deletes a snapshot of a filesystem. | Pause, Resume |
| FSCK | Checks metadata integrity. | Pause, Resume, Abort |
| DATA\_REDUCTION | Reduces data. | Pause, Resume, Abort |
| DATA\_DEFRAG | Defragments data. | Pause, Resume, Abort |
| FILESYSTEM\_RECOMPRESS | Rewrites the data of a filesystem to compress or decompress it. | Pause, Resume, Abort |
| QUOTA\_COLORING | Applies or clears a directory quota on a directory that already contains data. Requires an active Data Services container. | Pause, Resume, Abort |
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.
Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:
```
GET https://docs.weka.io/4.4/operation-guide/background-tasks.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.