Mastodon

Author: henry@warren.io

Categories
Release notes

Warren Cloud Platform v2.101 – multilingual UI, billable IPv4, acccess Grafana via admin panel

New Features:

  • Translation support for UI, emails, and PDF invoice templates.
  • Ability to charge users for assigned floating IP addresses.
  • Admin users can now manage (view, add, and delete) other admin users.
  • Virtual network DHCP now offers Cloudflare DNS as default option.
  • Onboarding page can be configured to allow users without a payment method.
  • Link to Grafana added to Admin Settings page.

Improvements:

  • Handling of error cases when adding a credit card fails has been improved.
  • Unassigned IP addresses with price 0 are no longer shown on invoices and usage reports.

Bug Fixes:

  • Fixed bug preventing to create VM from an existing disk.
  • User cannot delete their billing account anymore if they have a DB service on that account.
  • Fixed issue with error handling on the onboarding page.
  • Admin: Fixed issue when DB service was deleted, the underlying VMs did not get properly cleaned up.
  • Admin: Fixed initial prices view when no prices have been set yet.
  • Admin: VM migrate modal now fits longer host names.
  • Admin: Paying with Omise payment method now creates proper Google Analytics events.

Translation Support

Warren Cloud Platform now supports translation for the UI, emails, and PDF invoice templates. This means that users can now choose to view the UI in their preferred language, and emails and invoices will be sent in the language of the user’s account.

Language preferences are being asked from the end user during the signup.

Floating IP Address Pricing

It is now possible to charge users for assigned floating IP addresses. This can be useful for providers who want to offer additional services to their customers, or for organizations who want to track the usage of floating IPv4 addresses.

To set the price for floating IP addresses, go to the Billing section of your account settings and enter the desired price. The price will be applied to all active resources with public (IPv4) addresses.

Admin User Management

Admin users can now manage (view, add, and delete) other admin users. This can be useful for organizations with multiple administrators, or for providers who want to give their customers more control over their accounts.

To manage admin users, go to the Admin Users section of your account settings. You can add new admin users by entering their email addresses and selecting the roles you want to assign them. You can also delete existing admin users by clicking the Delete button next to their names.

Cloudflare DNS as Default Option for Virtual Network DHCP

Virtual network DHCP now offers Cloudflare DNS as default option. This can improve the performance and reliability of DNS resolution for virtual machines.

To enable Cloudflare DNS, go to the Virtual Networks section of your account settings and select the network you want to configure. Then, click the Edit button and select Cloudflare DNS from the DNS Provider dropdown menu.

Onboarding Page Configuration

It is now possible to configure the onboarding page for post-payment flow. This means that you can choose to allow users to enter the platform without adding a payment method.

Link to Grafana Added to Admin Settings Page

A link to Grafana has been added to the Admin Settings page. Grafana is a popular open source data visualization and monitoring tool.

To access Grafana, click the Grafana link on the Admin Settings page. You will be redirected to the Grafana login page.

Bug Fixes

A number of other bugs have been fixed in this release, including:

  • Fixed bug preventing to create VM from an existing disk.
  • User cannot delete their billing account anymore if they have a DB service on that account.
  • Fixed issue with error handling on the onboarding page.
  • Admin: Fixed issue when DB service was deleted, the underlying VMs did not get properly cleaned up.
  • Admin: Fixed initial prices view when no prices have been set yet.
  • Admin: VM migrate modal now fits longer host names.
  • Admin: Paying with Omise payment method now creates proper Google Analytics events.

We hope you enjoy this new release of Warren Cloud Platform! Any questions and feedback always welcome.

Categories
Uncategorized

Host Pools

Page
edited by
Sven-Erik Petermann

Brief Introduction

Compute hosts (aka hypervisors) can be grouped to different host pools. This allows separation of hosts with different purpose and configuration. Virtual Machines designated to a particular pool will stay in the pool throughout its lifetime. Only admin users have the capability to move Virtual Machines between pools. Host pools are grouped by their storage cluster. Virtual Machines can only be moved between pools that are using the same storage cluster.

Host pools are designed to separate physical hosts that still share common network.

Possible Use Cases

Selling different CPU types for different prices, user can choose between standard Intel (cheaper) or performance AMD (more expensive) CPU when creating a Virtual Machine.

Selling hosts with different resource overcommit with different prices, user can choose between standard cluster with CPU overcommit or “dedicated” cluster where the number of vCPUs running on a host never exceeds the actual number of CPU threads the host has.

Easier management of hosts with different CPU features.

Selling block storage from clusters with different performance for different prices.

Keeping good players in a hidden pool on hosts where no other user can create resources.

Keeping bad players in a hidden pool on hosts where they don’t bother other users.

Host Pool Concepts

Storage pool corresponds to a Ceph cluster. There can be multiple Ceph clusters in one location. A compute host can only be configured to one storage pool/Ceph cluster.

Host pool is a group of similar hosts (hypervisors). When starting a Virtual Machine, a host is chosen from the VM’s designated host pool. CPU type, CPU and RAM allocation coefficients and the storage pool are defined on host pool level. To end-user, host pool is shown as server class when creating a new VM.

Pool Management

Every location has at least 1 host pool and storage pool defined by default. Host pools are local to a location, they cannot span across multiple locations. All hypervisors must be assigned to some host pool.

Host pool properties, shown on Host pools admin view :

Name of the host pool and UUID for filtering in hosts table

Name can be seen by end-user

Description about host pool purpose

Can be seen by end-user

CPU model – this is QEMU / KVM CPU model configuration that applies to all hosts in that pool.
The CPU model configuration in QEMU/KVM refers to the settings that define the virtual CPU that is presented to the guest operating system running inside a virtual machine. Selecting a CPU model that closely matches the hardware of the physical host machine can help ensure compatibility with device drivers and system libraries.
It is vital to configure a CPU model that is supported by all hosts in that pool. The platform expects virtual machine live migrate to be possible between all hosts in a particular pool. A common baseline between host CPUs must be configured if hosts with different CPUs are present in one pool.

Virtual CPU coefficient – CPU overcommit factor for virtual CPUs. The number of available vCPUs on a host is calculated by the number of physical threads on that host times the host pool CPU overcommit factor. E.g., for a host that has 24 physical threads in a host pool that has 2.5 CPU coef., the number of vCPUs available for virtual machines will be
24 physical threads * 2.5 = 60 virtual CPUs

Virtual Memory coefficient – Defines how much physical memory is reserved for allocation to virtual machines. The recommended value 0.8 means that 80% is reserved for VMs and the other 20% is reserved for management overhead. E.g., for a host that has 140 GB of RAM in a host pool that has 0.8 Mem coef., the amount of RAM available for virtual machines will be
140 GB * 0.8 = 112

Default Pool for VMs – new VMs that are created will have this as their designated pool by default. The default applies to both API and, if pool (server class) selection is visible to the user, also to the default selection in Create VM view.

Pool is visible – A visible pool is presented to the end user in Create VM view as a selectable Server class option. The Default Pool for VMs will be auto-selected. By default host pools are not visible.

Hosts – shows number of hosts assigned to that host pool. Clicking on the number will open Hosts table view, showing only hosts from that pool.

Default pool for hosts – useful when adding new hosts to a location, new hosts will self-register to the selected pool.

Storage pool name and UUID – all hosts in this pool must all have this storage cluster configured in their OS configuration. Different host pools can share a storage pool. Storage pool name is not visible to end-user.

Action buttons for host pools –

Edit (similar to Create view)
Storage pool can only be changed if there are no hosts in that host pool currently. When a storage pool is selected for a host pool, only hosts with the same storage pool can be added to the host pool.

Delete –

Host pool must not contain any hosts

No Virtual Machine must have this pool as its designated pool

Host Pools and Virtual Machines

All Virtual Machines have a designated host pool. When Starting a VM, a host is chosen from the VM’s designated pool only. If the pool does not have a host with enough resources available, the VM can not be started.

If no host pools have “visible” flag set then all newly created Virtual Machines will have the Default pool for VMs as their designated pool.

Visible host pools are presented to the user when creating a Virtual Machine as Server class option, in which case the selected class/pool will be the designated host pool for that VM.

End user is not allowed to change the designated host pool for an existing Virtual Machine. If such need arises it must be handled by platform admin.

Admins can change VM’s designated pool in VM details view. Do note that for a running VM, changing the designated pool does not cause the VM to switch hosts automatically.

VM disks are bound to the storage cluster that they were created in. This means that new designated pool selection is limited to host pools that have the same storage pool configured as the current designated pool. VMs cannot be migrated to another storage cluster.

After changing the designated pool, one of two actions should follow:

Manual Migrate of VM to a host in the designated pool. The platform will do a best effort of live migrating the VM. This will move the VM to a new host in the new pool, provided that the CPU model that was configured for the VM at the time of Start in the old pool is compatible with the new host in the new pool.

Use platform Stop and Start to have the VM running on a host from the designated pool. It is not enough to do reboot from inside the VM, this will keep the VM running on the same host as it was before, so Current pool will not be changed.
Admin can either Stop and Start from UI (or via API) by themselves, or ask the user to do so at a time most appropriate for their service.

Pricing for Different Server Classes

It is possible to set different prices for CPU, RAM and disk/block storage depending on the Server class (host pool). In the Prices configuration view, clicking on Add button (1) will open a dialog with dropdown (2) from where either a location or a host pool in that location can be chosen. Locations or host pools that already have a price list are not selectable in the dropdown.

Although the host pool specific price list will only apply to CPU, RAM and disk, it is advisable to define prices for all resources, copying them from the location specific or the global default price list. This will make it easier for users to read and understand price lists.

Running Virtual Machines are priced by their Current pool, so if the designated pool is changed, but the VM is not Stopped and Started or Migrated, it will be still priced by their current pool. This in accordance with user VM details view, where Server class will show Current pool for running VMs, no matter what their designated pool is.

Categories
Uncategorized

2023 Release Notes

Page
edited by
Jevgeni Võssotski

October

20.10.2023
12.10.2023

September

26.09.2023
06.09.2023

August

21.08.2023

July

18.07.2023
03.07.2023

June

12.06.2023

April

13.04.2023

March

27.03.2023
1.03.2023

February

09.02.2023
01.02.2023

January

19.01.2023
09.01.2023

October

20.10.2023

v2.100.1

FIXED

Fixed VM cloning regression bug

More resilient sending of invoice/report emails, fixing issues with sending process hanging indefinitely

Fixed an issue with parallel creation of multiple network resources in Mikrotik backed SDN

12.10.2023

v2.100

NEW

Support for multiple Ceph clusters as different storage pools in one location

IMPROVED

Admin: VM Image edit view has now information about logo image format, including previews for different cases

FIXED

Fixed a loophole where user was able to delete their billing account while a VM was being created

Fixed invoice email sending, it will not be retried indefinitely if sending fails

September

26.09.2023

v2.99

NEW

Credit top up payments by Stripe can be configured to include a payment fee, see https://warrenio.atlassian.net/wiki/spaces/WARP/pages/46563694/Billing+and+Payment+System+Overview#Payment-Gateway-Fees for more information

IMPROVED

Put a limit to API gateway memory hogging issue

Admin: Direct URL to GraphQL Apollo Studio has been added to Admin setting area

FIXED

Admin: Fixed filtering Credit Notes by subtotal and other amounts in Reports/Invoices table

Admin: Simplified loader animation in admin tables to reduce browser CPU usage

Admin: Unnecessary New button was removed from admin area list view

06.09.2023

v2.98

NEW

Replaced the banner on user dashboard to promote dark theme feature

IMPROVED

Admin: Billing accounts now have Deleted At date

Admin: Windows price packages now show information about license price being included

Admin: Windows package prices are now shown in admin pricing area even if license price is 0

Admin: Greatly improved Hosts data loading time

FIXED

Fixed Download PDF for invoice/report

Admin: Button for scheduling new prices is now always visible, it was hidden in some specific cases

Admin: Fixed Delete VM in VM table action menu, now working for all locations

August

21.08.2023

v2.97

NEW

When adding a card with Stripe processor, user can be asked for their card address postal code to be used in Stripe rules

Platform support for MFA with Auth0. If switched on, all users who log in with username-password can enroll to MFA from their profile page. Additional Auth0 fees may apply if many users decide to turn on MFA for their account

Dark theme with automatic light/dark selection

Admin: It is now possible to schedule new prices for future and view historical prices

Admin: Billing account table has action button to delete an account and create a credit invoice, if needed, see Deleting billing account by administrator for more information

Admin: Price configuration view shows VM packages in a similar fashion to what the user sees

Admin: Support ticket can be opened from admin view either by a button at the end of sidebar list or by link in “URLs to know” section

IMPROVED

Optimised email sending during end of month report generation

FIXED

Server class prices are shown correctly to end user in setups with no selectable locations

Floating IP’s billing account is now updated while updating linked resource’s billing account (VM, Services, Load Balancer)

Admin: Fixed report/invoice ID field in Monthly reports table, it now has both copy and open detail view buttons

July

18.07.2023

v2.96

NEW

Admin: Restoring deleted disks from Ceph trash is now possible with admin API, for more information, see Restoring a deleted Disk

FIXED

Nightly backups run in a more reliable way, reducing the chance of interruptions and backups not being created.

Admin: fixed a regression where usage reports generated for prepayment accounts could bypass crediting and result in positive total amount.

03.07.2023

v2.95

NEW

It is now possible to have custom VM packages that are shown in VM Create and other similar views

Admin: IP address action log query in admin API is much more flexible now, log can be queried by not just IP address UUID but also address or by VM or LB that the address was assigned to, see admin API documentation for more information

FIXED

Fixed the flow of deleting billing accounts, payment will be properly attempted for unpaid invoices

When deleting billing accounts, usage is either shown on an invoice or not shown in total usage, there can be no inbetween cases anymore

Better handling of deleting snapshots

Stabilised the background cleaning up process of IP addresses and Load Balancers

Prevented some S3 storage users to have duplicate entries on Warren side, causing confusion

Admin: Fixed VM Details view for deleted VMs, action buttons are now disabled

Admin: Refreshing VM Details view now updates all the data of the VM

June

12.06.2023

v2.94

NEW

Admin: Preliminary support for closing billing accounts and cancelling their debt with credit invoices

Admin: Virtual Machines table can now show Deleted VMs as well

Admin: Hosts table entries have a button to quickly show all VMs that are assigned to that host

IMPROVED

Cheapest VM option is always visible in Create VM view, if minimum CPU is bigger than 1 then cheapest option is shown as not available

More information is stored about user Consent for Terms of Service, handling ToS versions is now easier

Load Balancer API now has endpoint for getting one LB by its UUID

Migrated Google Tag Manager to Google Analytics 4

More flexible configuration for disk trashing on Ceph

In VM Create view, an appropriate warning is shown when user chooses root as username, some OSes might have an issue with that

Improved data synchronisation reliability in multi-location setups

Admin: View and Copy icons are now both available for IDs in tables

Admin: Host pool memory allocation coefficient upper limit is now 95% of host’s total memory

Admin: Suspended flag and reason can now be changed also for deleted billing accounts

FIXED

Fixed the issue of updating S3 bucket billing account, in some cases it had no effect

Cloning a VM with a discontinued OS name and version now creates the clone as expected

Not showing Windows license price information in VM creation details when license price is 0

Fixed failover selection resetting bug in Managed Service Create view

Fixed currency formatting in some views

Fixed deleting IP addresses of Terminated billing accounts

Admin: Fixed VM detail view first loading of action log

Admin: When a new VM image conf is added with an existing name then an appropriate error message is shown, creating a duplicate entry is prevented

Admin: Fixed a few Host pool table usability issues

April

13.04.2023

v2.93

NEW

UI can be configured to show a link to an external support web page

Admin: Rescue Mode for admin now accepts custom ISO name to boot from, enabling admins to recover VMs with any ISO they prefer

IMPROVED

Virtual Machine disk trashing is now configurable with two new parameters: how long will disks be kept in trash, and a threshold for bypassing trash if a deleted disk is only few hours old

In Virtual Machine list with more than 10 running VMs, informative notice is shown to explain that monitoring data is loaded only for the first 10 VMs, for performance reasons

Admin: Hosts can now be marked as deleted to hide them from admin views, setting the flag is possible via admin API

FIXED

Fixed the issue of unnecessarily splitting VM compute usage on report

March

27.03.2023

v2.92

NEW

User can choose Server class with different CPU and RAM configuration when creating a Virtual Machine, different classes have different pricing options, read more from Host Pools documentation

Extended support for Omise payment options, credit card processing with Omise is now available

IMPROVED

Buttons on horizontal menu bars now have a minimal no-text form to fit more buttons on a narrow screen, e.g. on mobile devices

UI texts regarding initial free credit and referral bonus amounts have been clarified

Admin: Alert is raised in Slack when a control node or hypervisor disappears

Admin: Admin dashboard has location selection that takes up less space to nicely accommodate many locations

Admin: Migrate dialog shows loading animation if host and pool information is not yet loaded

Admin: Added a more convenient button to open details view from admin tables – a button is added as the first column, where applicable

FIXED

Fixed a race condition bug where it was possible for one Public Floating IP address to be allocated to multiple Virtual Machines at once

Fixed an issue that prevented multi-node DB service to acquire Floating IP at creation stage

Duitku payment methods properly handle the case when no credit card type processor is enabled

Admin: Admin tables’ global search now properly works for integers

Admin: Fixed IP Addresses table filter for address, e.g. entering 208 as address filter now returns only addresses that actually contain 208

Admin: Informative price packages in price settings view are not selectable anymore

Admin: Virtual Machine details view for admins now properly refreshes data after migration

1.03.2023

v2.91

NEW

VM is now assigned to a server class (Host Pool) that administrators can manage. This is also a milestone for further development to allow users choosing a server class for their VMs.

Added support for Omise payment gateway

VM details view has now a reboot action. This force stops and then starts the VM.

IMPROVED

VM resize modal and similar modals now show VM name next to UUID for clarity

When user has at least one suspended billing account then user is not allowed to create any new billing accounts anymore

Deleted VM disks are placed in trash area for two weeks period, during which they can be recovered by administrators

Admin: VM details view toolbar now includes all available actions and has a scrollbar for smaller screens

Admin: Calculated prices for VM packages can now be observed on the price list page while updating unit prices

Admin: Referral program settings can now be managed from the Admin settings page

FIXED

Fix delay in syncing delegated access data that resulted in restricting accessible accounts incorrectly

Creating Windows VM using existing disk as a boot disk will not induce file system checks from now on

February

09.02.2023

v2.90.1

NEW

Support for Stripe on-session payments to enable users in India to manually top up their accounts

FIXED

Adding, removing, attaching and detaching a disk now properly lock the VM to prevent parallel processes from breaking VM configuration

Fixed bug of parsing bucket metadata timestamps for some versions of of Ceph Object Storage

01.02.2023

v2.90

NEW

Admin: Host pool management is now visible to admins. Some host related parameters can be configured on pool level. This is a preliminary step to provide different resource pools to end users in a future release.

IMPROVED

Virtual Machine Reinstall allows now to choose a different OS version to install, E.g. Ubuntu 20.04 could be overwritten with Fedora 37

Improved the error message that is shown to end user if Virtual Machine Start fails because no resources are available

User interface now has a HTML meta tag to prevent robots from indexing the page

Admin: Clicking Host UUID in Virtual Machine admin table now opens the respective Host in Hosts table

FIXED

Billing account capabilities are now properly checked when creating new network resources, suspended accounts cannot create Floating IP addresses anymore

Closed access to user interface server side context, malicious user could have used control plane compute resources for their own good

Fixed Intercom chat button to remain in the top left corner

Fixed the complimentary free credit amount on user Dashboard, in some cases the amount was not displayed properly

January

19.01.2023

v2.89

NEW

Stripe card payment for top ups can be configured to convert the payment amount to some other currency at the moment of payment; decision to convert is done based on billing account and card country

Billing account default VAT percentage can be configured per country, i.e. if a new billing account is created and a country is selected, then the VAT percentage that is configured for that country will be automatically set to the new account

09.01.2023

v2.88

NEW

Virtual Machine can now be Reinstalled, i.e. boot disk overwritten with a clean base OS image

Admin: Floating IP pools can be configured in such a way that some pools are only available when admin specifically requests for an IP address from that pool using Create Floating IP admin API

Admin: Floating IPs can now be unassigned from the IP Addresses table

Admin: Added improved Services table to admin for faster loading

IMPROVED

Collect additional customer data and send it to Stripe payment gateway to satisfy country specific regulatory requirements (like India)

If taking a snapshot fails because of missing or unresponsive QEMU Guest Agent then the error message will say so

FIXED

Fixed a bug that prevented assigning Floating IP to a managed service

Fixed a bug with cleaning up dangling network interfaces in Tungsten SDN

Create VM API rate limit now actually applies only to Create VM endpoint, VM Start and Stop can now be called much more frequently

Admin: Various minor fixes regarding IP Addresses table

Categories
Uncategorized

Network

Network - warren.io

Page
edited by
Rauno Väli

What is it?

A section on Warren platform for handling creation and management of network-related resources

Main Components:

VPC – Virtual Private Cloud

A virtual private cloud (VPC) is a private network interface for collections of your resources. VPC network is available at no additional cost. This provides a certain level of isolation between the different resources. At the moment, each user can create up to 10 VPC pools.

Floating IP

A Floating IP is an IP address that can be instantly moved from one VM to another VM in the same VPC network. This also enables quick response for failovers when some system components malfunction, lifting reliability of the whole system. Floating IP addresses allocated to resources are not priced in particular. There is no limit for end-user to create unallocated floating IP addresses, but each of them is separately priced.

Categories
Uncategorized

Warren.io Release Notes v2.99

Release Date: September 27, 2023

New Features

  • Credit top up payments by Stripe can be configured to include a payment fee. See Billing and Payment System Overview | Payment Gateway Fees: https://paymentcloudinc.com/blog/payment-gateway-fees/ for more information.

Improvements

  • Put a limit to API gateway memory hogging issue.
  • Admin: Direct URL to GraphQL Apollo Studio has been added to Admin setting area.

Bug Fixes

  • Admin: Fixed filtering Credit Notes by subtotal and other amounts in Reports/Invoices table.
  • Admin: Simplified loader animation in admin tables to reduce browser CPU usage.
  • Admin: Unnecessary New button was removed from admin area list view.

Additional Information

This release includes a number of new features and improvements, as well as bug fixes.

The most notable new feature is the ability to configure credit top up payments by Stripe to include a payment fee. This is a useful feature for businesses that want to pass on the cost of credit card processing to their customers.

The improvements in this release include a fix for a memory hogging issue in the API gateway, as well as the addition of a direct URL to GraphQL Apollo Studio in the Admin setting area. This will make it easier for developers to access and use the GraphQL API.

The bug fixes in this release include a fix for filtering Credit Notes by subtotal and other amounts in the Reports/Invoices table, a simplification of the loader animation in admin tables to reduce browser CPU usage, and the removal of an unnecessary New button from the admin area list view.

Categories
Uncategorized

Aligning Data Center Hardware to Your IaaS Business Plan with Warren.io

As data centers, web hosting and cloud providers grow their infrastructure to accommodate more business requirements, they are oftenforced to ask themselves whether the hardware of their chosen IaaS platform will be able to meet these evolving needs of their customers? WithWarren.io’s comprehensive IaaS solution that ensures IT services are always up-to-date with the best technology and performance metrics available, navigating this challenge has never been easier. The entire organization from your customers, application developers right through to system admins can rest assured that their data center hardware is always in line with a service plan designed around your specific business goalsand objectives. Let’s explore how Warren.io helps you align your hardware investments for optimal support of your IaaS requirements today!

Table of Contents

  1. Maximize Computing Efficiency: Leveraging Compute Resources and Host Pools for Different Product Tiers
  2. Planning out local network performance: routers, switches and NICs for all the physical nodes
  3. Selecting the right block storage devices for Ceph distributed storage
  4. Energy efficiency and sustainability is increasingly important for the years to come
  5. Free consultation from Warren team on planning your hardware setup

Maximize Computing Efficiency: Leveraging Compute Resources and Host Pools for Different Product Tiers

Compute Host Pools for different performance tiersUsually a good business advice would be to go for the highest amount of cores the money can buy. Your customers can be kept happy by keeping the physical hypervisor server “over commitment” setting of Warren cloud platform within the optimal bounds for your specific end user profile who runs on the specific server. The use of “Platinum” CPU features is not common for a general public and the clock speed doesn’t have to be the highest for most common workloads. Unless of course you are deliberately aiming high to include the “GHz rate” in your marketing and sales materials to communicate the added value of your better CPU performance and flat even ratio of your vCPU to physical CPU core.

Configurable physical CPU time that a single vCPU unit is actually allowed to deliver in a scope of the given server or any logical group of physical servers know to Warren users as “Host Pools”. This type of segmented or tiered offering described above is easily set up from the Warren interfaces available to the IaaS provider.

More detailed overview can be found from the Wiki page about Host Pools https://warrenio.atlassian.net/l/cp/Nm1w0tg0 in Warren documentation.

Planning out local network performance: routers, switches and NICs for all the physical nodes

Warren.io recommends using network devices that can allow at least 25Gb throughput for faster access to Ceph distributed storage and better overall network performance. This is because Ceph is a high-performance distributed storage system that can generate a lot of network traffic, especially when running demanding workloads.

Here are some examples of Warren.io network devices that can support a minimum of 25Gb throughput:

  • Switches:
    • Arista 7280E
    • Cisco Nexus 9300
    • Dell PowerSwitch 9500
  • Routers:
    • Cisco ASR 9000
    • Juniper MX10000
    • Nokia MX9600

It is important to note that the specific network devices that you choose will depend on your specific needs and requirements. For example, if you are running a large-scale Ceph cluster, you will need to use network devices that can support a high volume of traffic.

Here are some tips for improving your overall network performance:

  • Use redundant network links to avoid bottlenecks and downtime.
  • Use a load balancer to distribute traffic evenly across your network devices.
  • Use a traffic management solution to prioritize traffic and optimize performance.
  • Monitor your network performance regularly and make adjustments as needed.

By following these tips, you can ensure that your network is able to support the demands of Ceph and your other workloads.

Selecting the right block storage devices for Ceph distributed storage

The cost of a distributed nature is evident with Ceph. Ceph is a distributed storage system, which means that it replicates data across multiple nodes to achieve high availability and scalability. However, this replication also adds to the cost of Ceph.

The price of Ceph redundancy may be high, but it offers unparalleled flexibility. Ceph can be deployed on a wide variety of hardware, and it can be scaled to meet the needs of even the most demanding workloads. Ceph is also highly reliable, and it can tolerate the failure of multiple nodes without losing data.

Balancing read/write performance is crucial when designing a Ceph cluster. Ceph can be optimized for either read or write performance, but it is important to choose the right configuration for your workload. For example, if you are running a database application, you will need to optimize for write performance. If you are running a web hosting application, you will need to optimize for read performance.

Understanding the true cost of high IOPS is important when designing a Ceph cluster. High IOPS performance can be achieved by using high-performance hardware, such as SSDs and NVMe drives. However, this hardware can be expensive. It is important to consider the cost of high IOPS performance against the actual requirements of your customers’ workloads.

Here are some tips for reducing the cost of high IOPS in Ceph:

  • ol]:!pt-0 [&>ol]:!pb-0 [&>ul]:!pt-0 [&>ul]:!pb-0″ value=”3″>Use Ceph features such as SSD journals and cache tiering to improve IOPS performance without using expensive hardware for all of your data.
  • It is important to note that there is no one-size-fits-all solution for reducing the cost of high IOPS in Ceph. The best approach will vary depending on your specific workload and budget constraints.

    With the recently deployed Warren release it is now also possible to allocate different tier Storage Pools for the previously available compute Host Pools. 

    Energy efficiency and sustainability is increasingly important for the years to come

    It is important to consider the energy consumption of every single device inside of your logical chassis when choosing the required stack of infrastructure devices for Warren.io based cloud cluster. This is because energy costs can be a significant expense for data centers, and reducing energy consumption can also help to reduce greenhouse gas emissions.

    Here are some tips for choosing energy-efficient infrastructure devices:

    • Look for devices with Energy Star certification. Energy Star is a government-backed program that certifies products that meet high energy efficiency standards.
    • Choose devices that are right-sized for your needs. Oversized devices will consume more energy than necessary, even when they are not fully utilized.
    • Consider using renewable energy sources to power your data center. This can help to reduce your reliance on fossil fuels and lower your carbon footprint.

    Here are some examples of energy-efficient infrastructure devices that are suitable for Warren.io cloud platform:

    • Servers:
      • Dell PowerEdge R760xd
      • HP ProLiant DL380 Gen10 Plus
      • Lenovo ThinkSystem SR650
    • Storage:
      • Dell PowerStore 5500
      • HPE Nimble Storage HF450
      • Pure Storage FlashArray//X
    • Networking:
      • Arista 7280E
      • Cisco Nexus 9300
      • Dell PowerSwitch 9500

    These devices are all Energy Star certified and offer a variety of features to help reduce energy consumption, such as power management modes and intelligent cooling.

    By choosing energy-efficient infrastructure devices, you can help to reduce your data center’s energy costs and environmental impact. This is important for both financial and sustainability reasons.

    In addition to the above, you can also implement the following measures to reduce the energy consumption of your Warren.io cloud platform:

    • Use thin virtualization/containerization to consolidate servers and reduce the number of physical devices that you need to operate.
    • Use energy-efficient cooling systems, such as free cooling or water-based cooling.
    • Implement data center infrastructure management (DCIM) software to monitor and optimize your energy consumption.

    By taking these steps, you can make a significant impact on the energy efficiency of your Warren.io cloud cluster. Stay tuned for more news on what Warren platform will be able to deliver in terms of optimizing your data center energy bill while making your operation more sustainable as a byproduct.

    Free consultation from Warren team on planning your hardware setup

    Get in touch to get your hardware sourcing plans right from the beginning and have a piece of mind about the projected return on investment.

Categories
Uncategorized

Cloud Computing Platforms: Alternative To OpenStack and Proxmox

Open source, despite popular misconception, does not imply that it is entirely free of cost. While open source software is often available at no monetary expense, it signifies a different ideology – the freedom to access, modify, and distribute the source code. 

Proxmox

Proxmox VE is a more user-friendly and accessible solution for smaller to medium-sized environments. It is easy to set up and manage, and it provides a comprehensive set of features for managing virtual machines and containers. Proxmox VE also includes its own built-in storage solution, which makes it a good choice for organizations that want to avoid the complexity of managing multiple storage backends.

  • Limited support: Proxmox VE is an open-source project, so there is no single vendor that provides support for the entire platform. This can make it difficult to get help when you need it.
  • Stability issues: Proxmox VE has been known to have stability issues, especially in large-scale deployments.
  • Security vulnerabilities: Proxmox VE has been known to have security vulnerabilities, which have been exploited by attackers.
  • Performance bottlenecks: Proxmox VE can be complex to optimize for performance. Organizations have experienced performance bottlenecks due to poorly configured components or inadequate hardware.

OpenStack

OpenStack is a more powerful and scalable solution for large-scale cloud deployments. It offers a wide range of features for managing virtual machines, containers, storage, networking, and more. OpenStack is also highly customizable, which allows organizations to tailor it to their specific needs. However, OpenStack can be complex to set up and manage, and it requires specialized expertise.

  • Complexity: OpenStack’s many components require correct configuration and management, which can be a challenge for organizations lacking expertise.
  • Cost: Large-scale OpenStack deployments can be expensive due to specialized hardware, software, and support requirements.
  • Security: OpenStack’s complexity exposes potential vulnerabilities. Proper configuration and security measures are vital for data and application protection.
  • Lack of support: As an open-source project, obtaining comprehensive support can be challenging.
  • Fragmentation: OpenStack’s size and complexity contribute to fragmentation within the community, making resource and support discovery difficult.

Here is a more detailed comparison of Proxmox VE, OpenStack and Warren platforms:

Proxmox VEOpenStackWarren
Ease of useEasierMore complexEasy to use by design
ScalabilityGoodExcellentExcellent
CustomizabilityLimitedHighCustomized selectively by Warren
FeaturesComprehensive set for managing VMs and containersWide range of features for managing VMs, containers, storage, networking, and moreWide range of IaaS end user features. Unique set of features for effortless IaaS service delivery.
SupportActive community support and paid enterprise support optionsActive community support and paid support options from vendors and partnersManaged Service, Slack based support from core developers
TCO for IaaS deliveryAffordableExpensiveAffordable


Warren

Warren.io is a turnkey cloud platform that can be used to offer a public cloud service. It includes everything from advanced IaaS features to hourly billing and day-to-day operations. This makes it a good choice for organizations that want to offer a public cloud service without having to invest in the infrastructure and expertise required to build and manage their own cloud platform.
 
Here are some of the benefits of using Warren.io to offer a public cloud service:
 
  • Turnkey solution: Warren.io is a complete solution that includes everything you need to offer a public cloud service, from the infrastructure to the billing and management system. This saves you the time and effort of having to build and manage your own cloud cloud platform and integrations like for instance billing and metering.
  • Advanced IaaS features: Warren.io includes a wide range of advanced IaaS features, such as virtual machines, containers, block and object storage, networking and load balancing. This gives your customers the flexibility to choose the resources they need and to scale their applications up or down as needed.
  • Hourly billing: Warren.io uses hourly billing, which makes it easy for your customers to pay for the resources they use. This is a more transparent and cost-effective billing model than traditional upfront pricing models of most “old school” hosting providers who charge you a month up front.
  • Day-to-day operations: Warren.io platform prioritizes of all the day-to-day operations of your cloud service, such as easy live migration of provisioned resources, monitoring performance, and resolving issues. This frees you up to focus on other aspects of your business.

Which one is right for you?

If you are looking for a user-friendly and accessible virtualization management solution for a smaller to medium-sized environment, then Proxmox VE is a good choice. If you need a more powerful and scalable solution for a large-scale cloud deployment, then OpenStack is a better option.

Here are some additional things to consider:

  • Your budget: Proxmox VE is a free and open-source solution, while OpenStack can be more expensive, depending on the support options you choose.
  • Your skill level: Proxmox VE is easier to set up and manage than OpenStack, but it still requires some technical expertise. If you need a solution that is easy to use and maintain, then Proxmox VE is a good choice.
  • Your specific needs: If you have specific requirements for virtualization, storage, networking, or other features, then you should carefully compare the features of Proxmox VE and OpenStack to see which one better meets your needs.

Ultimately, the best way to decide which virtualization management platform is right for you is to carefully consider your specific needs and requirements.

Open source, despite popular misconception, does not imply that it is entirely free of cost. While open source software is often available at no monetary expense, it signifies a different ideology – the freedom to access, modify, and distribute the source code. 

Proxmox

Proxmox VE is a more user-friendly and accessible solution for smaller to medium-sized environments. It is easy to set up and manage, and it provides a comprehensive set of features for managing virtual machines and containers. Proxmox VE also includes its own built-in storage solution, which makes it a good choice for organizations that want to avoid the complexity of managing multiple storage backends.

  • Limited support: Proxmox VE is an open-source project, so there is no single vendor that provides support for the entire platform. This can make it difficult to get help when you need it.
  • Stability issues: Proxmox VE has been known to have stability issues, especially in large-scale deployments.
  • Security vulnerabilities: Proxmox VE has been known to have security vulnerabilities, which have been exploited by attackers.
  • Performance bottlenecks: Proxmox VE can be complex to optimize for performance. Organizations have experienced performance bottlenecks due to poorly configured components or inadequate hardware.

OpenStack

OpenStack is a more powerful and scalable solution for large-scale cloud deployments. It offers a wide range of features for managing virtual machines, containers, storage, networking, and more. OpenStack is also highly customizable, which allows organizations to tailor it to their specific needs. However, OpenStack can be complex to set up and manage, and it requires specialized expertise.

  • Complexity: OpenStack’s many components require correct configuration and management, which can be a challenge for organizations lacking expertise.
  • Cost: Large-scale OpenStack deployments can be expensive due to specialized hardware, software, and support requirements.
  • Security: OpenStack’s complexity exposes potential vulnerabilities. Proper configuration and security measures are vital for data and application protection.
  • Lack of support: As an open-source project, obtaining comprehensive support can be challenging.
  • Fragmentation: OpenStack’s size and complexity contribute to fragmentation within the community, making resource and support discovery difficult.

Here is a more detailed comparison of Proxmox VE, OpenStack and Warren platforms:

Proxmox VEOpenStackWarren
Ease of useEasierMore complexEasy to use by design
ScalabilityGoodExcellentExcellent
CustomizabilityLimitedHighCustomized selectively by Warren
FeaturesComprehensive set for managing VMs and containersWide range of features for managing VMs, containers, storage, networking, and moreWide range of IaaS end user features. Unique set of features for effortless IaaS service delivery.
SupportActive community support and paid enterprise support optionsActive community support and paid support options from vendors and partnersManaged Service, Slack based support from core developers
TCO for IaaS deliveryAffordableExpensiveAffordable


Warren

Warren.io is a turnkey cloud platform that can be used to offer a public cloud service. It includes everything from advanced IaaS features to hourly billing and day-to-day operations. This makes it a good choice for organizations that want to offer a public cloud service without having to invest in the infrastructure and expertise required to build and manage their own cloud platform.
 
Here are some of the benefits of using Warren.io to offer a public cloud service:
 
  • Turnkey solution: Warren.io is a complete solution that includes everything you need to offer a public cloud service, from the infrastructure to the billing and management system. This saves you the time and effort of having to build and manage your own cloud cloud platform and integrations like for instance billing and metering.
  • Advanced IaaS features: Warren.io includes a wide range of advanced IaaS features, such as virtual machines, containers, block and object storage, networking and load balancing. This gives your customers the flexibility to choose the resources they need and to scale their applications up or down as needed.
  • Hourly billing: Warren.io uses hourly billing, which makes it easy for your customers to pay for the resources they use. This is a more transparent and cost-effective billing model than traditional upfront pricing models of most “old school” hosting providers who charge you a month up front.
  • Day-to-day operations: Warren.io platform prioritizes of all the day-to-day operations of your cloud service, such as easy live migration of provisioned resources, monitoring performance, and resolving issues. This frees you up to focus on other aspects of your business.

Which one is right for you?

If you are looking for a user-friendly and accessible virtualization management solution for a smaller to medium-sized environment, then Proxmox VE is a good choice. If you need a more powerful and scalable solution for a large-scale cloud deployment, then OpenStack is a better option.

Here are some additional things to consider:

  • Your budget: Proxmox VE is a free and open-source solution, while OpenStack can be more expensive, depending on the support options you choose.
  • Your skill level: Proxmox VE is easier to set up and manage than OpenStack, but it still requires some technical expertise. If you need a solution that is easy to use and maintain, then Proxmox VE is a good choice.
  • Your specific needs: If you have specific requirements for virtualization, storage, networking, or other features, then you should carefully compare the features of Proxmox VE and OpenStack to see which one better meets your needs.

Ultimately, the best way to decide which virtualization management platform is right for you is to carefully consider your specific needs and requirements.

Categories
Uncategorized

Starting out as a Cloud Service Provider (CSP)

Understanding Warren partnership

Warren’s offering consists of a software platform and accompanying managed services provided by an experienced team based in Europe, Estonia.

Introducing Warren platform

Warren is a complete self-service cloud platform for regional infrastructure service providers. Functionally rich, technologically modern and architecturally reliable turnkey cloud commercialisation solution packaged as a service and driven by your success to compete in the global landscape.

Warren services provided

The cloud hosting platform of Warren is offered as a managed service.

That means that Warren’s team deploys the Warren platform on your hardware in your data center and keeps it up and running and up to date.

Technical support team

The technical support is carried out via a support ticketing system and a dedicated Slack channel. The Slack channel is used for the day to day technical support where Warren developers are always available to support, make amends (if necessary) and offer advice however they can.

The Warren technical support team operates during working hours (GMT+3).

Around the clock tailored support is available as a part of a separate SLA contract.

Continuous platform upgrades

Warren delivers planned bi-monthly deployments of software upgrades to deliver new features and maintain the existing functionality. The upgrade deployments do not involve any downtime or maintenance window. This keeps the platform up to date and keeps the software delivery process very flexible and robust.

Feature requests

Warren and the team accept feature requests from the partnering cloud providers. Moreover, we expect our partners to provide constant feedback on both the platform and accompanying services alike.

Sign up for a 30 minute product demo or Q&A session
Product
 
Company
Support
  • Warren is a high availability and decentralized hybrid cloud management platform based on sophisticated proprietary machine learning algorithms that makes application production much more coherent, synoptic, efficient and affordable. Warren platform is supported by European Regional Development Fund in the sum of 301 500 EUR (07.08.2017 - 31.12.2018).

General terms & conditions | Privacy policy