Linux servers are everywhere. Linux OSes run in most hypervisors in private and public cloud. Linux runs in physical clusters in private cloud, and VMs in private, public, hybrid cloud. According to Gartner’s estimates for 2015,
- Linux revenue grew to $2 Billion. It’s more than UNIX revenue at $1.8 Billion.
- For every 3 Windows servers, there are 2 Linux servers.
Enterprises run 100s of critical applications with databases like Oracle, DB2, MySQL, PostgreSQL, and NoSQL on Linux OS. It’s not uncommon to find multi-TB databases and filesystems in Linux systems. Enterprises want to make sure that these applications are not only high performance and secure, but are also resilient. In other words, their data is backed up and retained to meet their compliance needs, and in the event of a disaster these applications can be recovered very quickly with minimum downtime.
The three most important considerations for resiliency are the three I’s:
- Incremental forever backup – Full backups impact production application performance. So it’s important that the backup product be able to perform incremental forever backups to minimize the impact on production apps and reduce backup window by upto 10x
- Instant recovery – Whether the application running on Linux OS has multi TB database, or a large file system with millions of files, the backup product should be able to do instant recovery in minutes. Low RTO & low RPO is a must for critical applications.
- Infinite retention – Enterprises should have the flexibility retain their backups for hours, days, weeks, months, years, or even decades, if they need to. Yet they should still have the ability to offer instant access to data, whether it is days old or decades old.
Actifio satisfies all these requirements as shown in Fig 1.
Customer Case Study – Integrated Trading Company
Here is an example of how a customer benefited with Actifio solution for large DB2 databases running in Linux OS.
For each Linux VM or physical server, Actifio deploys a connector (like an agent). The connector has the brains to keep track of what blocks are changing in a volume (Fig 2).
The administrator defines SLAs that dictate when Actifio Sky has to initiate a backup, for example – every hour, every 6 hours, or every 24 hours. Based on those SLAs, Actifio coordinates with the connector to backup just the changed blocks. Before the changed blocks are backed up, the Actifio connector invokes pre- and post-scripts to freeze and then thaw the database so that the changed blocks that are captured are in a database-consistent mode.
After each incremental backup is done, Actifio synthesizes a virtual full backup (as shown in Fig 3). Thus, when the user has to do a recovery at a certain point in time, he always gets a full synthetic image.
Once the backups are done, based on SLAs, Actifio Sky replicates deduped or compressed backup images to a remote Actifio Sky running at a DR site or in a public cloud, such as Microsoft Azure, AWS or Oracle Cloud.
Both at the primary site and remote site, Actifio Sky can deliver instant recovery by mounting backup images to the recovery server with a few clicks (Fig 4).
Actifio Sky can retain data for the desired retention period according to the admin-specified SLAs. Sky can retain data for days, weeks, or months in its deduplication pool. For multiple years and decades, Actifio Sky can vault data to object storage platforms – in the cloud (e.g. Amazon S3, Azure Blob, Softlayer or Google Cloud Nearline), or on-premises (e.g. IBM Cleversafe / Cloud Object Storage). Actifio Sky can provide instant access to data stored in object storage, again by simply mounting an image, thus enabling enterprises to leverage the data for on-demand cloud analytics or data warehousing needs.
Test Data Management for test dev & DevOps teams
Most large enterprises have Application Test and Development or DevOps teams that need copies of production application data in their test environments. Actifio Sky enables those teams and DBAs to ‘instantly’ provision virtual copies off backup images to multiple Linux test environments in a self service manner. For ex – as shown in Fig 5, multiple virtual copies of a 10 TB database can be provisioned instantly in various test environments in a self service manner by Dev, QA teams.
This enables Test dev and Devops teams to always test their code against virtual copies of production data on VMs or physical servers from day one of their sprint cycles. This helps them catch critical defects early in Dev & QA cycles and improve product quality, and at the same time accelerate their release cycles. Since the virtual copies do not consume any extra storage, this also reduces the infrastructure requirements significantly for IT teams. The on-demand virtual data provisioning also fits well with DevOps teams that want to leverage on-demand compute in cloud for test dev.
Actifio delivers a single enterprise grade platform to provide mission-critical Linux based applications the capabilities they need for Backup, Long-Term Data Retention, Disaster Recovery, and Test Data Management for DevOps.
Want to see a demo on how you can provision a multi-TB database instantly for DR or test/dev? Send an email to email@example.com.