Byte-level file replication vs block-level disk replication

Byte-level file replication vs block-level disk replication in a high availability cluster

Overview

This article explores the pros and cons of byte-level file replication vs block-level disk replication in a high availability cluster. We are looking at volume of replicated data, impact on application data organization, recovery time, simplicity of implementation.

The following comparative tables explain in detail the byte-level file replication implemented by SafeKit, a high availability software product.

What is byte-level file replication?

Byte-level file replication (like with SafeKit) means that only modifications inside files are replicated.

Synchronous replication is required in a high vailability cluster for having 0 data loss in case of failure. Asynchronous replication is for backup solutions.

The volume of replicated data is reduced to information modified by applications inside their files. No extra data is replicated.

There is no impact on data organization for an application. For instance, if an application has its data in the system disk, byte-level file replication is working.

Recovery time (RTO) in the event of a failover is reduced to the application restart time on the secondary server's replicated files.

Finally, the solution is very simple to configure as only the paths of directories to replicate are configured.

What is block-level disk replication?

Block-level disk replication (like with DRBD) means that only modifications inside a disk are replicated.

The volume of replicated data is not reduced to information modified by applications. Extra data are replicated like the meta data for managing the disk (list of free blocks, file system internal information).

There is a strong impact on the organization of application data. All data must be localized in the replicated disk. At least, it requires an application reconfiguration. Or, it is impossible if some data to replicate are in the system disk, because this disk must remain specific to each server.

The recovery time (RTO) increases with the file system recovery procedure on the replicated disk after a failover.

Finally, the solution is not easy to configure because skills are required to configure a special disk with a file system. Additionally, application skills are required to configure application data in the replicated disk.

Pros and cons of byte-level file replication vs block-level disk replication

Cluster with byte-level file replication	Cluster with block-level disk replication
Product
SafeKit on Windows and Linux	Disks replication products as DRBD
Application data organization
0 impact on application data organization with SafeKit. Just define directories to replicate in real-time. Even directories inside the system disk can be replicated.	Impact on application data organization. Special configuration of the application to put its data in a replicated disk. Data in the system disk cannot be replicated.
Data replication
Synchronous byte-level file replication. Replicates file modification operations generated by application activity No meta data are replicated. Only data modified in files are replicated, not entire files (byte-level file replication). Synchronous replication to avoid data loss on failure.	Replicate all data modified inside a replicated disk. Application data plus meta data are replicated. For instance, last access time on a file is replicated (last access time is modified each time the file is read).
Complexity of deployment
No - install a software on 2 servers	Yes - require specific IT skills to configure OS and replicated disk
Failover
Just restart the application on the second server.	Remount the file system on the replicated disk. Pass the recovery procedure on the file system. And then restart the application.
Failback
Automatic failback. Resynchronization of data on the secondary server without stopping the application on the primary server. No application failover while data are not resynchronized.	All products are not at the same level of features.
Quorum and split brain
Application executed on a single server after a network isolation (split brain). Coherency of data after a split brain. No need for a third machine or a quorum disk or a special heartbeat line for split brain. More information on heartbeat, failover and quorum	Require a special quorum disk or a third quorum server to manage split brain.
Suited for
Software editors which want to add a simple high availability option to their application	Enterprise with IT skills in clustering.