Storage

Our high performance cloud storages are designed with security, speed, availability, flexibility, and ease-of-use in mind. All data is encrypted at rest and stored on enterprise grade SSD disks. Our storage options range across multiple performance classes and storage types, and are able to fulfill even the highest I/O requirements.

Our offering includes the following storage options:

Rocket Storage #

Rocket Storages are backed by NVMe drives attached directly to the Server which boasts increased I/O bandwidth and lower latency. With a RAID 10 implementation over four disks within a compute node, they are not replicated like Distributed Storage over multiple storage servers, therefore it is recommended to have a backup plan to recover your data if anything were to happen.

Distributed Storage #

Your data is replicated across three Affinity Zones within a location allowing our autorecovery and failover mechanisms to make sure that your data will always be available. This is achieved via networking and has certain trade-offs like lower I/O, bandwidth and higher latency. Distributed Storage comes in three performance classes: Standard, High and Insane. See the table below for more information and comparison to our Rocket Storage.

Object Storage #

Our S3-compatible Object Storage offering is developed keeping data sovereignty in mind, designed to be accessed from anywhere including websites, mobile applications or IoT devices. You can read more about our Object Storage offering here.

Storage Performance Classes #

Connectivity #

Each storage can only have one connected server.

Limits #

Rocket Storage currently has a snapshot capacity of one, with four backup schedules, and eight backups per schedule. See servers for limits on connecting storages to servers. All other limits can be lifted, these are static.

Templates #

We offer multiple OS templates {%gs-only%} on gridscale {%/gs-only%} like CentOS, Ubuntu, Windows etc.

Please note: Even though you can technically create very large storages, sometimes the underlying operating system cannot take advantage of that, like Windows with the standard MBR partition.

Backups #

With the Backup Center, a backup schedule can be created to create full image backups. Within the Backup Center interface, you can configure the default backup schedule for all new storage within a project.

Alternatively any snapshot created can be exported to Object Storage, effectively making it a backup.

Boot Device #

The first storage connected to the server becomes the boot device. It is possible that a server has no boot device even if multiple storages are connected, this normally happens when removing the boot device after connecting other storage. Simply navigate to the server in the expert panel to configure the boot device.

Windows Licencing and ISO Images #

We provide Windows templates and we also handle the licensing. It is not currently supported that you use your own licence on our infrastructure by using an ISO Image.

Hybrid Core Partners with their own license agreements, can choose to pay the licenses themselves for Windows servers running on their infrastructure.

FAQs #

I don’t see my preferred operating system listed. Can I upload my own ISO? #

Yes, if you enter a URL into our ISO Creator, we will download the corresponding ISO image for you and make it available within your account.

How can I delete a storage? #

Click on the item you want to delete and drag and drop it to the trash can that appears as soon as you click on it.

Is there a size limit for storages? #

The storage limits for your account can be viewed while trying to create a storage. However, our architecture also allows you to create storages with several terabytes. If you need to adjust your personal limits, just contact us.

Can I use other storages on a server besides the operating system storage? #

You can currently attach up to eight storages to a cloud server. A new storage without a template behaves like an empty hard disk. Before you can use the SSDs, a filesystem must be created and partitioned if necessary.

Can I assign servers, storages, snapshots, etc. to another user? #

No. This is not yet possible, but is planned for the near future.

How does the storage system work? #

The Distributed Storage is built using enterprise SSD storage from brands such as Intel or Samsung.

All data on Distributed Storage is replicated to at least three independently operating storage servers. Our storage works independently of the server host systems, so that a highly available data storage is also guaranteed, if a server host system has failed. Your storages are always available at any data center location. We don’t want to distribute the data across the boundaries of a data center, so you can always decide 100% where your data is stored.

Rocket Storage consists of NVMe drives attached directly to the Server which boasts increased I/O bandwidth and lower latency.

This interface, which is optimized for SSDs, allows the full performance potential of SSD technology to be exploited. The direct physical connection via PCI Express, the resulting elimination of the need for protocol translators and parallel processing of requests result in lower overhead. All of these factors increase the I/O performance of our Rocket Storage by up to a hundred times compared to ordinary storage, depending on the deployment scenario.

Our flexible S3-compatible Object Storage allows you to store and retrieve any amount data, designed to be accessed from anywhere including websites, mobile applications or IoT devices.

Stability and access control are covered through easy-to-use security features.

How much temporary “ephemeral” storage do I have per server? #

Our servers have no temporary or ephemeral storage. Every storage you create is always persistent block storage.

Do I always have to link a server to a storage? #

No, only if you want to boot the server from an installed template (this should be the normal case). Of course you can also use a server without local storage, for example by Netboot via PXE. This is possible because you can operate all services in your private networks, DHCP, PXE, VLANs.

How do I increase the size of my drive on the fly? #

You can adjust the storage space within the panel settings. In order for the new size to be usable, the filesystem has to be re-partitioned and enlarged.

{%gs-only%} You can do the same by following our tutorial. {%/gs-only%}

Can I use a storage in multiple servers at the same time? #

Technically we have already implemented this function, but so far we lack the concrete purpose for it.

Most file systems (EXT3/4, ReiserFS, BTRFS, NTFS etc.) are immediately damaged if they are integrated several times into one or differentoperating systems.

To prevent accidental data loss, we have decided to leave the function deactivated for the time being. If you have a specific application or scenario for this function, then feel free to contact us.

Can I use other storage than your SSD storage for my servers – e.g. for backup or archiving? #

Depending on your use case, you can opt for backup, automation and scheduling options within our Backup Center.

You can also perform instantaneous backups in real-time by exporting your snapshots to our S3-Compatible Object Storage.

{%gs-only%} In addition to supporting other third-party backup solutions such as Acronis, specifically for our Platform Services we’d recommend using Backup Ninja in order to store backups locally or on our S3-compatible Object Storage. {%/gs-only%}

When can I take a snapshot of my storage? #

You always have the possibility to create a new snapshot. It doesn’t matter if your storage is connected to a server or if the server has just started.

Please note that your file system may not be properly mapped in a snapshot when the connected server is running and data is read/written from your storage.

What do I have to consider with your Object Storage? #

You can use or configure the gridscale Object Storage via the Panel.

Our Object Storage is S3-compatible, you only need to enter the host name through gos3.io and connect you to Key and Secret. After successfully connecting, you can see your buckets.

What else can I do with snapshots? #

The snapshot function allows you to roll back a storage to an earlier state. You can also create a new, custom template from a snapshot and use it to install your servers.

This allows you to create individual server templates that allow you to start new, pre-configured systems within seconds. You can also export your snapshots to the Object Storage in case of backups or archives.

Which tool can you recommend for uploading to Object Storage? #

If you want to upload larger-sized data onto your Object Storage, we recommend using Cyberduck (Mac and Windows) or s3cmd (Mac and Linux).

How can I restore a snapshot? #

When rolling back a storage to the previous state, it is important that you first shut down the server.

Then click on the affected storage and simply select the desired snapshot. Request the rollback and in a few seconds the storage is reset to the previous state. Please keep in mind that you will lose all changes made after the snapshot was taken.

Which virtualization software and storage do you use? #

We use KVM as our Hypervisor and our storage offering backend consists of a Distributed SSD Storage and NVMe-based local Rocket Storage directly connected to the server. Our provisioning stack is an in-house development, as this was the only way to meet our high standards of quality, performance and availability.