Google Regional Persistent Disk
Are you seeking for a natively Google Compute Engine solution that offers 0% RPO, is production-ready, and provides high availability for your mission-critical workloads? You only need to look at Google Regional Persistent Disk. This blog post from Google Cloud explores how you can provide resilience, ease of management, and continuous protection for your most critical and demanding applications by utilizing Regional PD’s cross-zone synchronous replication capabilities.
Regional Persistent Disks
Regional PD: A mission well suited for it-important tasks
Workloads that are crucial to the daily operations of every modern firm are mission-critical. These are the systems that the company needs to function, thus any interruption to their workloads could have a big effect on the company.
Because of this, companies require highly available infrastructure, which includes highly available storage capable of:
- Replicate data seamlessly while maximising performance for certain kinds of tasks.
- Be extremely robust against a variety of problems that could impact availability.
- Be really easy to assemble and operate
For many of the mission-critical high availability storage requirements of Google Cloud customers, Regional PD is a strong fit and meets all of these criteria.
Incredibly effective for your workloads that are mission-critical
Google Cloud customers use Regional PD for MySQL, Postgres, SQL Server, Jupyterlab, Kafka, Druid, Redis, Solr, EventStoreDB, Elasticsearch, MongoDB, and other workloads due to its write and read performance of 1000MB/s and 80K IOPS per volume, respectively.
Extremely reliable and resilient to failure
With regional PD, synchronous replication between zones is continuous and has an RPO of zero in the event that one zone goes down. Regional PD prioritizes availability and automatically reroutes I/O to the last available storage replica in the event of a zonal outage. Regional PD self-heals and brings the unavailable replica back online to continue RPO=0 replication later, once the unavailable zone has recovered. To ensure workload availability, a virtual machine (VM) can be tied to the regional PD if it needs to be recovered in the second zone.
For your mission-critical applications, full replication means that all the data required to maintain the high availability of your compute instance is replicated. This includes not just data disc replication but also boot disc replication for virtual machine instances. When Virtual Machines (VMs) employ Regional PD as its boot disc, users are shielded against potential storage problems that could prevent the VM from booting and from data loss from VMs that rely on the boot disc for data, such as Windows VMs.
Easy to set up and maintain
Configuring Regional PD for mission-critical workloads is really simple. It only takes a few minutes to add a new Regional PD disc whether starting a new virtual machine or updating an existing one. Using the Google Cloud console, gcloud, Terraform, and REST APIs, you can add Regional PDs. Once configured, your workload can start serving reads and writes to the Google Regional Persistent Disk in the same way as it would with any other block storage, with the exception that all writes will automatically use synchronization replication to replicate to another replica on a secondary zone.
High Availability
Regional PD in Google Cloud services with high availability
A number of Google Cloud services, including as GKE Stateful High Availability (HA) Controller, Cloud SQL HA configuration, and Cloud Shell, leverage Regional PD in addition to Compute Engine workload deployments.
The completely managed database service offered by Google Cloud for MySQL, PostgreSQL, and SQL Server is called Google Cloud SQL. When enabled, as soon as a Cloud SQL database instance is created, the configuration immediately configures it to use Regional PD as the primary storage and replicates writes to a secondary zone. When necessary, failover is a straightforward one-step command that switches all networking, compute, and client applications to the secondary zone’s alternate instance.
The terminal window on Google Cloud called Cloud Shell gives you command line access to virtual machines (VMs) in your projects. Users get 5GB of Regional PD disk storage for their terminal’s home directory with each Cloud Shell terminal session. In the event that there are issues with the zone housing the home directory storage, this helps guarantee that the home directory is highly available. The Cloud Shell service automatically uses the extra Regional PD replica in the event of a zonal issue, ensuring uninterrupted access to home directory data.
Final reflections
This blog post from Google Cloud explains how Google Regional Persistent Disk was designed with the goal of maximizing availability for your mission-critical workloads in Google Cloud while also delivering ease in management, performance, and resilience. Furthermore, Google Cloud talked about how users have implemented Regional PD in expansive production settings and how it offers storage availability across a number of core Google Cloud services.
Concerning the local persistent disk
You can use Compute Engine’s Regional Persistent Disk storage option to create high availability (HA) services. Google Regional Persistent Disk provides high availability (HA) for disk data for up to one zonal failure by synchronously replicating data between two zones within the same region.
Workloads with a reduced Recovery Point Objective (RPO) and Recovery Time Objective (RTO) are intended for use with Regional Persistent Disk volumes.
- Replication of zones on discs for local persistent discs Regional Persistent Disk volumes hold disc data in primary and secondary zones.
- For disc attachment, the primary zone is the same as the VM instance.
- A secondary zone is a different area inside the same region of your choice.
In each of these zones, Compute Engine keeps copies of your local persistent disk volume. To guarantee HA, Compute Engine synchronously duplicates data you write to your disc to the disc replicas in both zones. To ensure longevity, the data of each zonal replica is distributed among several physical machines in the zone. Zone replicas guard against brief disruptions in one of the disk zones and guarantee that the Persistent Disk volume’s data is always accessible.
