Storage Topology

Path: Left sidebar > Advanced > Storage Topology

When to Use:

• After host or OSD topology changes.

• Before creating, editing, or deleting CRUSH rules or erasure-code profiles.

Purpose:

This page explains how to validate placement topology and manage the policies that pools use for replica or chunk placement.

Steps:

1. Open Advanced > Storage Topology.

2. Review the CRUSH tree and confirm the current topology is correct.

3. Move to the CRUSH Rules or Erasure Code Profiles section for the target policy change.

4. Validate the new rule or profile before using it in a pool workflow.

Expected Outcome:

• You can confirm the cluster placement model and safely manage the policies that depend on it.

What You See:

• The CRUSH tree, CRUSH rule list, erasure-code profile list, and the create/edit/delete workflows for those policies.

What This Screenshot Shows:

• The screenshots on this page show the topology tree and the policy forms used to manage CRUSH rules and erasure-code profiles.

Actions in This Screen:

• Expand and review the CRUSH tree.

• Create, edit, or delete CRUSH rules.

• Create or delete erasure-code profiles.

If this fails:

1. Do not continue if the topology tree already looks incorrect.

2. Confirm no dependent pool is using the rule or profile you plan to remove.

3. Re-validate placement behavior in Pools after the topology change.

Storage Topology

The Storage Topology page combines three admin sections on one scrollable page:

• CRUSH Tree: visual hierarchy of data-placement topology.

• CRUSH Rules: placement policies pools use for replica/chunk mapping.

• Erasure Code Profiles: redundancy profiles for erasure-coded pools.

This page represents the CRUSH map (Controlled Replication Under Scalable Hashing), which Ceph uses to decide which OSDs store each piece of data.

Note

Values and names shown in screenshots are UI references from one environment. Use values that match your own topology and policy.

CRUSH Tree

The CRUSH Tree is an interactive hierarchy showing how Ceph organizes infrastructure for placement decisions.

What This Screenshot Shows: Storage Topology Dashboard (UI Reference; Values Depend On Your Environment).

Purpose:

• To verify host/OSD placement before creating rules and pools.

When to Use:

• After adding new hosts or OSDs.

• Before creating/editing CRUSH rules.

CRUSH Tree Node Types

Node Type	Badge Label	What It Represents
Root	root	Top-level bucket where placement starts
Region / Datacenter	region / datacenter	Optional geographic grouping level above racks/hosts
Rack	rack	Physical rack grouping
Host	host	Individual server node
OSD	osd + class/status	Leaf disk where objects are stored

Reference Topology Example (From Screenshot)

default [root]
\- dc-us-va-Sterling-karios-staging-01 [datacenter]
\- rack-r87 [rack]
|- i9-storage [host] -- 4 items
| |- osd.2 [osd][ssd] up 0.233 TiB
| |- osd.4 [osd][nvme] up 0.233 TiB
| |- osd.5 [osd][nvme] up 0.233 TiB
| \- osd.6 [osd][hdd] up 0.910 TiB
\- siva-ceph [host] -- 3 items
|- osd.0 [osd][hdd] up 0.437 TiB
\- osd.1 [osd][hdd] up 0.910 TiB

How To Read OSD Leaf Entries

Item	Meaning
First badge	osd confirms this is an OSD node
Second badge	Device class (ssd, nvme, hdd)
Status indicator	up means daemon is healthy and online
Right-side size	Raw capacity of that OSD
Parent item count	Number of direct children under non-leaf nodes

How To Use The CRUSH Tree

1. Open Advanced > Storage Topology.

2. Expand/collapse nodes with chevron >.

3. Verify hosts appear under the correct rack/datacenter buckets.

4. Verify OSD device-class labels are correct.

5. Verify OSD status is up; investigate deviations via Infrastructure > OSDs.

Expected Outcome:

• You confirm topology and class mapping are correct for rule-based placement.

Tip

Always validate the CRUSH tree after host/OSD expansion before creating new pools or assigning class-specific rules.

TOOLTIP - Failure Domain And CRUSH Tree:

• host failure domain separates replicas across hosts.

• rack failure domain separates replicas across racks.

• osd failure domain separates at disk level only.

CRUSH Rules

CRUSH Rules define data-placement behavior for pools. Every pool references one CRUSH rule.

Purpose:

• To verify which placement policy each pool is using.

• To validate failure-domain and media-class design before pool changes.

When to Use:

• Before creating a new pool.

• Before editing an existing pool’s placement policy.

CRUSH Rules List - Column Reference

Column	What It Shows
ID	Internal numeric rule identifier
Name	Rule name (for example replicated_rule, nvme-rule)
Type	replicated or erasure
Failure Domain	Replica/chunk separation level (for example host, osd)
Used By	Pools currently using this rule
Actions	Create, Edit, Delete (Edit/Delete only when no pools use the rule)

CRUSH Rule Actions

Action	Description
Create	Define a new rule with Root, Failure Domain, and optional Device Class
Edit	Change failure domain or device class for an unused rule
Delete	Remove an unused rule

Reference Rules In UI (Example)

ID	Name	Type	Failure Domain	Used By
0	replicated_rule	replicated	host	default
1	nvme-rule	replicated	osd	default-nvme
2	ssd-rule	replicated	osd	default-ssd
3	hdd-rule	replicated	osd	default-hdd

TOOLTIP - replicated_rule:

• host failure domain is the safer default for multi-host resilience.

TOOLTIP - Class Rules:

• nvme-rule targets NVMe OSDs.

• ssd-rule targets SSD OSDs.

• hdd-rule targets HDD OSDs.

How To Create A CRUSH Rule

Path: Advanced > Storage Topology > CRUSH Rules > + Create

What This Screenshot Shows: Create CRUSH Rule Panel (UI Reference; Values Depend On Your Environment).

Purpose:

• To create explicit placement behavior for new or retuned pools.

When to Use:

• Before creating a pool or before changing pool placement policy.

Steps:

1. Scroll to CRUSH Rules section.

2. Click top-right + Create.

3. Enter Name.

4. Select Root.

5. Select Failure Domain.

6. Select Device Class (or None for mixed media).

7. Click Create.

Expected Outcome:

• Rule appears in the CRUSH Rules list with a new ID.

• Rule becomes selectable in Storage > Pools.

Create CRUSH Rule - Field Reference

Field	Value / Options	Description
Name *	Text	Unique descriptive rule name
Root	Dropdown (for example default)	Top-level CRUSH bucket
Failure Domain	Dropdown (for example host, osd)	Replica/chunk separation boundary
Device Class	None / ssd / nvme / hdd	Media filter for tiered placement

Warning

Delete a CRUSH rule only when Used By is empty. Deleting an in-use rule can cause pool placement issues.

Tip

Use class-specific rules for tiering: NVMe for hot workloads, SSD for mid-tier, HDD for capacity.

How To Edit A CRUSH Rule

Path: Advanced > Storage Topology > CRUSH Rules > edit action

Purpose:

• To update placement behavior for a rule that is not in use.

When to Use:

• Before assigning the rule to pools.

• When you need to correct failure domain or device-class targeting.

Steps:

1. In CRUSH Rules, identify the rule to modify.

2. Confirm the rule is not used by any pool (Used By is empty).

3. Open the edit action for that rule.

4. Update Failure Domain and/or Device Class.

5. Save the rule.

Expected Outcome:

• The rule definition updates immediately.

• The updated rule is available for future pool assignment.

Warning

Do not modify placement policy while recovery is active. Confirm all PGs are active+clean before topology/rule changes.

How To Delete A CRUSH Rule

Path: Advanced > Storage Topology > CRUSH Rules > trash icon

Purpose:

• To remove an obsolete rule that is no longer used by any pool.

When to Use:

• After migrating all dependent pools to a different CRUSH rule.

Steps:

1. In CRUSH Rules, find the target rule.

2. Confirm Used By is empty.

3. Click the trash icon in Actions.

4. Confirm deletion.

Expected Outcome:

• The rule is removed from the list.

• The rule is no longer available for pool assignment.

Warning

Never delete a rule that is still used by a pool. Reassign dependent pools first, then delete.

Erasure Code Profiles

Erasure Code Profiles define EC chunking and parity behavior for erasure-coded pools.

Purpose:

• To define redundancy efficiency for erasure-coded pools.

• To confirm which profile exists before pool creation.

When to Use:

• Before creating erasure-coded pools.

• During resilience/capacity tuning for object workloads.

What This Screenshot Shows: Erasure Code Profiles Section (UI Reference; Values Depend On Your Environment).

Erasure Code Profiles - Column Reference

Column	What It Shows
Name	Profile identifier
Plugin	EC plugin (for example jerasure, isa, lrc)
k (data)	Number of data chunks
m (coding)	Number of parity chunks and failure tolerance
Actions	Create or Delete (Delete only when profile is unused)

Profile Actions

Action	Description
Create	Define k, m, plugin, and optional device-class targeting
Delete	Remove an unused profile

Reference Profiles In UI (Example)

Name	Plugin	k (data)	m (coding)
test	jerasure	2	1
default	jerasure	2	2

TOOLTIP - Understanding k And m:

• Objects are split into k data chunks plus m parity chunks.

• Profile survives up to m simultaneous OSD failures.

• Storage overhead is (k+m)/k.

Example:

• k=4, m=2 stores chunks across 6 OSDs, survives 2 OSD failures, and has about 1.5x overhead (vs 3x for 3-replica).

How To Create An Erasure Code Profile

Path: Advanced > Storage Topology > Erasure Code Profiles > + Create

What This Screenshot Shows: Create Erasure Code Profile Panel (UI Reference; Values Depend On Your Environment).

Purpose:

• To define efficiency and failure-tolerance model for erasure-coded pools.

When to Use:

• Before creating an erasure-coded pool.

Steps:

1. Scroll to Erasure Code Profiles section.

2. Click top-right + Create.

3. Enter Name.

4. Select Plugin.

5. Set Data chunks (k).

6. Set Coding chunks (m).

7. Set Crush failure domain.

8. Set Technique.

9. Set Packetsize.

10. Set Crush root and Crush device class.

11. Click Create EC Profile.

Expected Outcome:

• Profile appears in list immediately.

• Profile becomes available in erasure-coded pool creation workflow.

Create EC Profile - Field Reference

Field	Value / Options	Description
Name *	Text	Unique profile name
Plugin	Dropdown (for example jerasure)	EC algorithm
Data chunks (k) *	Number	Data chunk count
Coding chunks (m) *	Number	Parity chunk count
Crush failure domain	Dropdown	Chunk separation boundary
Technique	Dropdown (for example reed_sol_van)	Mathematical encoding method
Packetsize	Number	Internal EC packet size
Crush root	Dropdown	Root placement bucket
Crush device class	Dropdown	Class target (or automatic selection)
Directory	Read-only path	EC plugin library location

Warning

EC profiles cannot be modified after creation. Create a new profile when k or m requirements change.

Tip

Use replicated pools for simpler operations unless EC efficiency is required for large, capacity-focused datasets.

How To Delete An Erasure Code Profile

Path: Advanced > Storage Topology > Erasure Code Profiles > trash icon

Purpose:

• To remove an unused EC profile that is no longer needed.

When to Use:

• After confirming no pool references the profile.

Steps:

1. In Erasure Code Profiles, locate the profile.

2. Confirm no erasure-coded pool uses this profile.

3. Click the trash icon in Actions.

4. Confirm deletion.

Expected Outcome:

• The profile is removed from the list.

• It cannot be selected for new pool creation.

Warning

EC profiles in use by pools cannot be deleted. Remove dependent pools first.

Troubleshooting - Storage Topology

Problem You See	Most Likely Cause	What To Do
Host or OSD missing in CRUSH Tree	Host/OSD not registered or deployed	Verify Infrastructure > Hosts and Infrastructure > OSDs
OSD appears under wrong host	Hostname/bucket mapping issue at provisioning	Escalate for CRUSH bucket correction
Cannot delete CRUSH rule	Rule still referenced by a pool	Reassign pool to different rule, then delete
Cannot delete EC profile	Profile still referenced by EC pool	Remove dependent pool first
New rule missing in pool dropdown	UI not refreshed	Refresh Pools page and reopen create/edit flow
OSD class appears incorrect	Provisioning class detection mismatch	Recheck class assignment with administrator

Note

If issues persist, raise a support ticket via Monitoring > Alerts or the Karios Support.