Pools

Path: Left sidebar > Storage > Pools

When to Use:

• Before creating file systems, block images, or object workloads.

• When you need to create, validate, edit, or safely delete a pool.

Purpose:

This page explains the end-to-end pool workflow, including first-pool creation, validation, tuning, and retirement.

Steps:

1. Open Storage > Pools.

2. Review existing pool health and PG state.

3. Use + Create Pool or the row actions for the target pool.

4. Validate the pool tabs before attaching workloads to it.

Expected Outcome:

• You can create or modify a pool and confirm it is healthy enough for dependent DFS features.

What You See:

• Pool list rows, grid/list toggles, create and row actions, and expanded tabs for details, performance, and configuration.

What This Screenshot Shows:

• The screenshots on this page show the main Pools landing page, the create-pool workflow, and the expanded validation tabs.

Actions in This Screen:

• Create a pool.

• Expand a pool row to validate health and configuration.

• Edit or delete a pool from the row actions.

If this fails:

1. Return to Dashboard and confirm the cluster is healthy enough for pool changes.

2. Stop if PGs remain non-clean or raw capacity is already too high.

3. Resolve application-tag or placement issues before retrying the action.

A pool is the base storage unit in Karios DFS. You must have a ready pool before using File System, Block Images, or Object Storage buckets.

If you are new to DFS storage, this page gives you the exact sequence to create, validate, edit, and safely delete pools.

What a Pool Controls

A pool defines:

• how many copies of data are kept (replica behavior)

• where data is placed (CRUSH/failure-domain behavior)

• capacity policy (optional quotas)

• workload intent (application tag)

Pool Type Quick Choice

Pool Type	How It Works	Best For
Replicated	Stores N full copies across different disks/nodes.	Most workloads. Use this when unsure.
Erasure Coded	Splits data into chunks + parity.	Capacity-optimized large sequential workloads.

Tip

Recommended default for new users: Replicated with Replica Size = 3. You can use a different value if your environment policy requires it. Follow 2.2.1 Erasure Coded Pool Creation Path when your workload requires Type = Erasure Coded.

Before You Create Any Pool

Run this checklist first from Dashboard:

1. Cluster status is OK (or non-blocking WARN).

2. OSD values show matching up and in counts.

3. Monitor quorum is healthy.

4. Raw capacity is below 80% used.

If any check fails, resolve infrastructure health first.

Page Orientation (What You See)

On Storage > Pools page:

• Top-right: + Create Pool button

• Next to it: grid/list view toggle icons

• Table columns: Name, Data Protection, Applications, PGs, PG Status, Usage, Read/Write, Actions

• Per-row actions (⋮): Edit Pool and Delete Pool

• Per-row chevron (>): expands tabs Details, Performance, Configuration

Storage > Pools landing page (reference example values from a sample workspace).

Pool Actions - Why, Impact, And When To Act

Action	Why User Does This	Impact	When To Do It
Create Pool	Creates a dedicated storage target for a workload.	Adds a new placement domain with its own protection and quota policy.	Do this before creating file system, block image, or object workload data.
Expand Row And Validate Tabs	Confirms the pool is healthy and correctly configured.	Prevents attaching workloads to a misconfigured or non-clean pool.	Do this immediately after create, after edits, and during alerts.
Edit Pool	Aligns protection, quota, or compression settings to workload changes.	Can trigger temporary rebalancing and short non-clean PG transitions.	Do this when workload policy changes or capacity/protection alerts appear.
Delete Pool	Retires unused storage and returns capacity to cluster.	Permanently destroys all objects in that pool.	Do this only after all dependencies are removed and backups are confirmed.

Create Your First Pool (New-User Runbook)

Recommended Starter Profile

Use this as a reference profile for a safe first pool. Replace values with your environment and workload requirements.

Field	Recommended Value	Why
Type	Replicated	Broad compatibility and predictable behavior.
Application	cephfs (for File System) or rbd (for block)	Enables correct workload behavior/optimization.
Replica Size	3	Production-safe default resilience.
Min Size	2	Allows writes with one replica temporarily unavailable.
PG Autoscale	On	Lets cluster adjust PGs over time.
Compression	None initially	Avoids surprise CPU overhead until workload is known.

Step-by-Step Create Flow

1. Open Storage > Pools.

2. Click + Create Pool.

3. In Pool Name, enter your environment-specific name using lowercase + hyphen style (example only: cephfs-data-01).

4. In Type, select the option that matches your workload policy. Replicated is the recommended first-run option.

5. In Application, select one of:

   • cephfs For File System Metadata/Data Workloads

   • rbd For Block Image Workloads

   • rgw For Object Gateway Workloads

6. Set Replica Size based on your resilience requirement. If unsure, use 3 as the recommended starting value.

7. Set Min Size based on your write-availability policy. Common starting value is 2 when replica size is 3.

8. Keep PG Count at the system default unless platform guidance says otherwise.

9. Set PG Autoscale per policy. Recommended setting is On.

10. Keep Crush Rule at environment default unless architecture requires custom rule.

11. Set Compression Mode based on workload profile. None is a safe first-run choice.

12. Set Quotas only if you need hard limits now. Leave empty for no limit.

13. Review your values, then click Create Pool.

What This Screenshot Shows: Create Pool panel in a reference environment.

Important

Do not leave Application as None. Untagged pools trigger a WARN health state and make intent unclear for future operations.

Erasure Coded Pool Creation Path

Use this path when Type = Erasure Coded.

Prerequisite:

• At least one EC profile exists in Advanced > Storage Topology.

Steps:

1. Open Storage > Pools and click + Create Pool.

2. Set Type to Erasure Coded.

3. Select the required EC profile.

4. Set Application for intended workload (for example rgw or environment-specific policy value).

5. Review values and click Create Pool.

6. Monitor PG Status until pool reaches active+clean.

Expected outcome:

• Pool row appears with EC data protection profile and reaches active+clean state.

If this fails:

1. Confirm selected EC profile exists and is valid.

2. Recheck cluster health and OSD availability.

3. Retry creation after resolving profile or health errors.

What You Should See After Create

Immediately after submit:

• Pool Row Appears In The Table

• PG Status can show transitional states briefly

Normal convergence path:

• Transitional/Initial State -> active+clean

Target readiness:

• Pool Reaches active+clean Within About 1-2 Minutes

If not ready in 5+ minutes:

• Check Infrastructure > OSDs

• Check Monitoring > Alerts

First-Pool Pass/Fail Check

Pass:

• Row Is Visible

• Application Tag Is Correct

• PG Status is active+clean

• No New Critical Alert Appears

Fail:

• Pool Row Absent After Create

• Wrong Application Tag

• PG Status stuck non-clean > 5 minutes

• Cluster Moves To Blocking Error State

View and Interpret Pool Details

Expand Pool Row

1. Click the row chevron > on the target pool.

2. Confirm the row expands into a detail panel directly under that pool.

3. Use the expanded header chips (for example <PG Count> PGs, <Object Count> objects, and <Used Capacity> used) for a quick health summary. The numbers shown in screenshots are reference examples only.

4. Open tabs: Details, Performance, Configuration.

Expanded Row Field Guide

Use this table to understand each visible row field before opening tabs.

Field	What It Indicates	How A New User Should Use It
Name	Pool identifier used by dependent services	Verify this matches your intended workload name.
Data Protection	Replica/EC profile (for example replica:x3)	Confirm protection level before storing production data.
Applications	Pool application tag (cephfs, rbd, rgw, etc.)	Ensure tag matches the workload type you plan to run.
PGs	Placement Group count for the pool	Use as a capacity/placement indicator; keep autoscale enabled unless directed otherwise.
PG Status	Current placement/health state	Target state is active+clean before production use.
Usage	Space consumed by objects in the pool	Confirm usage trend aligns with expected data growth.
Read / Write	Current IO activity signal	Use for quick confirmation that workload IO is reaching the pool.
Actions (⋮)	Row action menu	Use for edit/delete operations only after validation checks.

Expanded Tabs Guide

• Details: Validate Identity, Protection, Application, And Capacity Fields.

• Performance: Validate Throughput/IOPS Signals During Active Workloads.

• Configuration: Review Internal Parameters; Modify Only With Platform Guidance.

Pool Details tab.

Details Tab (What to Validate)

Check these values first:

• Pool Name Matches Intended Workload

• Type And Replica Settings Are Correct

• Application Tag Is Correct

• Usage Is Reasonable For Expected Stage

Performance Tab (What to Validate)

Use as quick health/traffic signal:

• Read/Write Throughput Present When Workload Is Active

• IOPS/Throughput Not Pinned At Abnormal Zero During Known Traffic

Pool Performance tab.

Configuration Tab (What to Validate)

Use this tab to validate internal pool properties before and after edits. For new users, treat this as a verification screen, not a tuning screen.

Configuration Validation Checklist

1. Confirm the table loads with Property, Value, and Source columns.

2. Confirm there are no unexpected empty values for core properties.

3. Confirm Any Value You Changed In Edit Flow Is Reflected Correctly Here.

4. Confirm there are no unplanned overrides introduced since last known-good state.

5. Confirm post-change pool health still returns to active+clean in the row header.

How to Read the Columns

• Property: Internal Configuration Key Name.

• Value: Effective Runtime Value Currently Applied.

• Source: Where The Effective Value Came From In Config Precedence.

What To Do If Something Looks Wrong

1. Do not apply more changes immediately.

2. Capture the property name and current value.

3. Compare with your last planned change set.

4. Re-check PG Status and Monitoring > Alerts.

5. Escalate to platform administrator before further tuning.

Pool Configuration tab.

Edit Pool Safely

How to Open Edit

1. In pool row, click ⋮ under Actions.

2. Select Edit Pool.

Pool row actions menu.

Safe vs High-Impact Changes

safe (with normal caution):

• Name updates

• Quota values

• Compression mode updates

Higher impact (plan maintenance window):

• Replica/min size changes (can trigger rebalance/recovery)

• Autoscale behavior changes at scale

Expected behavior after save:

• Change Is Applied Quickly

• Short Temporary PG Activity Changes Can Appear

• Pool Should Return To active+clean

Note

Pool type cannot be changed after creation. To change type, create a new pool and migrate workload data.

Delete Pool (Data-Destructive)

Warning

Deleting a pool permanently destroys all data in that pool.

Dependency Check Before Delete

Before deleting, confirm no active dependencies:

• Storage > File System: Pool Is Not Used As Metadata/Data Pool

• Block Storage > Block Images: No Image Depends On This Pool

• Object Storage > Buckets: No Bucket Path Depends On This Pool

Delete Steps

1. Click row ⋮.

2. Click Delete Pool.

3. Read dependency warning.

4. Type exact pool name in confirmation field.

5. Confirm deletion.

After delete:

• Row Is Removed

• Raw Capacity Is Returned

• Short Temporary WARN During Cleanup Can Be Normal

Grid/List and Search Tips

• Use List view for operational checks (status columns visible).

• Use Grid view for high-level browsing.

• Use search box to quickly locate target pool by name.

Troubleshooting - Pools

Problem	Most Likely Cause	What to Do
Pool not visible after create	Create action failed or permission issue	Re-check role, retry create, review alerts
PG Status stuck non-clean > 5 minutes	OSD/network/placement issue	Check Infrastructure > OSDs and Monitoring > Alerts
Pool shows degraded	One or more OSDs unavailable	Recover down OSDs first
WARN after create	Missing/incorrect application tag	Edit pool and set correct application
Pool usage reaches limit	Quota reached or cluster full	Adjust quota or add capacity
Cannot edit/delete	Insufficient role permission	Request Operator/Administrator role

Note

If issue persists after checks, escalate using Monitoring > Alerts and your support workflow.