Block storage
Overview
Coming from another cloud?
▸AWS·EBS
This Quake AI feature maps to AWS’s EBS.
▸Azure·Managed Disks
Azure Managed Disks
- Azure Managed Disks are fully managed with automatic redundancy (3 replicas, 99.999% SLA); Cinder durability depends on backend.
- Predefined disk types (Ultra Disk, Premium SSD v2, Premium SSD, Standard SSD, Standard HDD) with fixed performance tiers; Cinder uses volume types with configurable QoS.
- Disks billed on provisioned size regardless of use; Cinder billing typically on provisioned size too but varies by provider.
▸DigitalOcean·Volumes
Volumes
- Uses proprietary /v2/volumes API instead of OpenStack Cinder /v3/{project_id}/volumes; actions like attach/detach/resize via undocumented POST /volumes/{id}/actions rather than /os-volume-attach etc.
- Enforces single-instance attachment only (one Droplet at a time); OpenStack supports multi-attach volumes.
- Volume names must be globally unique per region; OpenStack uses IDs primarily with optional display names.
- Size specified in GiB (binary) with 1 GiB to 16 TiB range; OpenStack typically GiB but flexible.
▸Google Cloud·Persistent Disk
Persistent Disk
- Uses Google Compute Engine REST API instead of OpenStack Cinder API.
- Offers multiple performance tiers (pd-standard HDD, pd-balanced SSD, pd-ssd, pd-extreme) with performance scaling linearly with provisioned size, unlike Cinder's backend-dependent performance.
- Supports regional disks replicated synchronously across 2 zones for higher availability (better than 99.9999% durability), while Cinder volumes are typically zonal unless using replication features.
- Online resize to increase size without detaching; no manual striping/RAID needed as GCP handles distribution automatically.
▸Hetzner·Volumes
This Quake AI feature maps to Hetzner’s Volumes.
Block storage
Block storage volumes are persistent NVMe-backed disks that you create, attach to compute instances, and format with any filesystem. A volume's lifecycle is independent of the instance that uses it: detach a volume and reattach it to a different VM, snapshot it for point-in-time recovery, or clone it to spin up an independent replica.
OpenStack Cinder backs block storage. For details on how Quake AI implements OpenStack, see How Quake AI uses OpenStack.
Start here#
Start with Create a block storage volume. You size and create a volume, attach it to a running instance, and format it before you write data to it.
How-to#
- How to Clone a Block Storage Volume: Create an independent copy of an existing volume. The clone is a new, fully writable volume. Changes to the clone do not affect the source, and vice versa. You can clone a source volume whether it is...
- How to Create a Block Storage Volume: Create a persistent block storage volume that you can attach to virtual machine instances. Volumes persist independently of instances; you can detach, reattach, and snapshot them at any time.
- How to Create a Volume Snapshot: Create a point-in-time snapshot of a block storage volume. Snapshots are space-efficient: they store only the difference from the volume's base state, and you can use them to create new volumes or...
- How to Extend a Block Storage Volume: Increase the capacity of an existing block storage volume. Volume extensions are irreversible: capacity can only grow. Quake AI's Cinder supports online extend on an attached, in-use volume when you...
- How to Schedule Volume Snapshots and Restore Data: Schedule recurring Cinder volume snapshots with cron on a control instance or a CI workflow. Set a retention period, remove expired snapshots, and test restores on a recovery volume.
- How to Transfer Volume Ownership Between Projects: Move a block storage volume from one Quake AI project to another. The source project creates a transfer record. The target project accepts the record with the transfer ID and authorization key.
Concepts#
Concepts
- Volumes: Applications need storage that survives reboots, instance deletions, and infrastructure changes. Volumes provide that: persistent block storage devices that attach to instances, behave like physical...
Related builds#
Templates
This validated OpenTofu template composes Compute, Network, and Block Storage into a self-hosted extract-and-load host you run on infrastructure you control.Custom package · ~$68/mo · 30 minutes
This pattern composes Compute, Network, and Block Storage into a self-hosted workflow orchestration host you run on infrastructure you control.Custom package · ~$34/mo · 25 minutes
This validated OpenTofu template composes Compute, Network, and Block Storage into a self-hosted business-intelligence host you run on infrastructure you control.Custom package · ~$67/mo · 25 minutes
This pattern composes Compute, Network, and Block Storage into a self-hosted API gateway on infrastructure you control.Custom package · ~$31/mo · 30 minutes
This pattern composes Compute, Network, and Block Storage into a self-hosted low-code platform for internal tools and admin panels, on infrastructure you control.Custom package · ~$32/mo · 20 minutes
This pattern composes Compute, Network, and Block Storage into a self-hosted scheduling and booking platform, on infrastructure you control.Custom package · ~$66/mo · 30 minutes
Deployments
Stand up a retrieval-augmented generation pipeline on Quake AI by composing Ollama inference with Qdrant vector search. You ingest documents, embed chunks with Ollama, store vectors in Qdrant, and...~$195/mo · Custom package · 30 minutes
Run the durable control plane on Quake AI CPU compute: a Redis or Valkey queue and a worker process accept jobs, submit each one to a third-party GPU or model API, poll for completion, retry on failure, and land results in Object Storage. The queue and workers stay always-on and cheap; the GPU work runs on a provider you choose.~$30/mo · Custom package · 45 minutes
Stand up a Harbor container registry on a private subnet with one floating IP for HTTPS access using the harbor-registry OpenTofu template.~$72/mo · Custom package · 40 minutes
Stand up a private Valkey or Redis cache on a Quake AI private network using the validated OpenTofu template redis-cache. You set a password and apply the template, confirm the cache binds to the...~$14/mo · Custom package · 30 minutes
Stand up a Varnish Cache HTTP accelerator on one Quake AI instance using the validated OpenTofu template edge-cache. You apply the template, launch a private origin on the same subnet, confirm a cache...~$14/mo · Custom package · 45 minutes
Stand up Forgejo with a bundled Forgejo Actions runner on one CPU VM using the forgejo-git-ci OpenTofu template.~$34/mo · Custom package · 35 minutes
Pricing data last validated: . For current rates, check quake.ai/pricing.
Reference and operations#
When a volume gets stuck in a transitional state, refuses to delete, or fills up, work through Volume troubleshooting.
Volumes attach to Compute instances. If you need bucket storage that you reach over an S3 API instead of a filesystem, see Object storage.
Quick answers
Common Block Storage tasks
See Also
Was this page helpful?