Machines protection consists of several steps:
- Installing agents and configuring replication
- Getting consistent machines snapshots
- Detecting changes on machines between two snapshots
- Sending deltas to DC, taking into account WAN optimization and deduplication
- Storing of incremental backups in DC, taking into account snapshot storage policies
Retrieving machines snapshots and detecting changes¶
Depending on initial virtualization platform, obtaining consistent snapshots of machines is possible in many ways:
- in case of replicating machines on VMware, getting snapshots of Windows and Linux machines is done by standard tools of VMware (VMware CBT API), where there is a chain of internal API calls and snapshots including the means of machines OS.
- for Windows machines getting consistent snapshots goes through Volume Shadow Storage (VSS) calls and getting changes between two Application-consistent snapshots
- for Linux machines, obtaining consistent snapshots goes through the implementation of the VSS snapshot analog for block devices and sending changes to the snapshot store.
In case of VMware infrastructure replication, VMware Tools must be installed on your virtual machines.
Sending deltas to storage. Data deduplication¶
Deltas are accumulated changes since the last replication.
Sending changes to the DC comes from customer agent secure HTTPS protocol. This is deduplication process, when customer application sends only those data that are not yet in storage for the given customer. This allows to significantly save the amount of transferred data, reduce load on network and at times accelerate replication of machines.
In addition to WAN deduplication, optimization includes compression of network traffic.
Creating a group of machines¶
After creating a customer dashboard during product initial configuration to replicate machines, go to his page and set machine groups. Grouping machines allows to merge them by functionality or common parameters of replication schedule.
To create a group click Add at the bottom of the Customer page.
It is possible to add an arbitrary number of groups. By default each group inherits global settings for replication schedule and snapshot retention policies for customer. Their change for group will be described below. Note that group name must be unique within the customer.
Each customer must have at least one group. The Default group which includes all machines that have not been added to other groups is created by default.
To move machines between groups, select them and click Move to group in the menu Actions. Note that if replication schedule and snapshot storage policies parameters were not redefined for machines, these parameters will be inherited from new group when transferred to another group.
Machines replication and distribution between groups¶
When groups are created go directly to replicating machines. To do this use machine replication process (Section: ACP - Machine Replication Process) directly from the menu or through the group settings and the menu Replicate new platform. The difference is that with the second option a group for the machine at the migration process step will be immediately selected.
Select replicated platform type.
Select a group where all machines replicated by this agent (later machines can be distributed between groups) will appear. The last step is to download the agent, which is ready to be installed on customer’s infrastructure.
Agent installation and configuration¶
VMware Agent requirements¶
Hystax VMware Replication Agent requires the following user permissions in vSphere(Role “VMware Consolidated Backup user” in vCenter):
- Virtual machine - Configuration – Disc Lease
- Virtual machine - Provisioning – Allow read-only disk access
- Virtual machine - Provisioning – Allow virtual machine files upload
- Virtual machine - Snapshot management – Create snapshot
- Virtual machine - Snapshot management – Remove Snapshot
Also permission to access CBT is necessary for correct application work. To enable CBT:
Virtual machine - Configuration – Disc change tracking
Ports for correct agent work:
- DR host - tcp/80, tcp/443
- vSphere host - tcp/443
- ESXi host(s) - tcp/udp/902
- Receive logs from the Acura cluster - udp/12201
Windows and Linux Agents requirements¶
Requirements for correct work of Windows agent: - Hardware:
- Memory: 2 GB RAM
- CPU: x64 processor
- Disk space: 100 MB required for product installation and not less than 15% free space of disk size for VSS snapshots creation
- Microsoft .NET Framework 4.0
- Requirements for correct work of Linux agent:
- Memory: 500 MB RAM
- Disk space: 100 MB required for product installation and not less than 15% free space of disk size for snapshots creation
Install Windows and Linux agents¶
To install Windows and Linux agents just download msi file for Windows and deb / rpm package for Linux and deploy them on guest OS. After successful installation, the agents will automatically register the machine in the system and it will appear in the machine list in the group selected on the previous step.
Install VMware agents¶
To install VMware agents download an .ova file and deploy one instance of the agent on each of the ESXi hosts of VMware vSphere. Note: VMware vSphere can issue a warning about the presence of an unknown configuration parameter for the virtual machine. The service parameter “hvragent” is specially added and it is not a security risk.
To deploy a new agent follow the instructions below:
- Use VMware Web Client / vSphere Client / VMware fusion.
- Select downloaded .ova template with the agent for deployment.
- Select the storage to which the agent will be uploaded.
- Select disk provisioning and network that have Internet access
- Start the agent when deployment process is complete
After starting machines with agents, machines on the same ESXi host recognized by agents will appear in the list of customer machines in selected group by default. The machines will be added in the ‘Not replicated’ state (to enable select which machines need to be replicated and which ones are not); to start replication select necessary machines and click Start replication in the menu Actions.
Agent logs can be found, clicking the button To console in agent UI, stored in /logs/hvragent.log
To get current network settings, use the file /var/run/network/current-settings
To change network settings (for example, if agent doesn’t get IP address by DHCP), edit the file /var/run/network/interfaces. When the settings are changed, reboot the agent.
Setting replication schedule¶
When machines appear in the list and replication is enabled for necessary machines, then it is necessary to set replication schedules and snapshot storage policies (Sections: ACP - Replication schedule and ACP - Snapshot policy settings).
Settings can be changed globally for the customer, for each individual group or for individual machines.
Settings are inherited in the following way:
- By default, global settings are applied to all groups and machines
- Editing group settings, new settings are applied to all (in the group or to be added later) machines with the default settings, otherwise the settings of specific machines will work if they are changed
- Editing machines settings, machines are operated by settings of specific machine, otherwise the group settings or global settings are applied
Completed these steps, the machines are replicated and will be replicated based on schedule and rules, set during the configuration.