This section is, strictly speaking, not a how to section but rather a huge checklist of all the other components one should consider in a complete backup/restore scenario for an organization's Universal Communications solution.
Lync does not stand alone in the infrastructure; it relies on many other components, and it would be fatal to neglect this. Restoring only Lync will not get you anywhere near a functional system.
In the final paragraphs, we'll try to summarize some of the other components we must pay attention to when considering the disaster recovery plan.
Let's talk about some of the dependencies in the infrastructure that are not covered by the procedures described so far. Just to make things clear, these are as important as the Lync infrastructure itself. In order to have a full and healthy backup, consider including the following elements described hereafter in your disaster recovery plan.
This section will not describe how to make backups or do restores of other systems, as it is beyond the scope of this book. It is here for reference and to give Lync administrator s a clearer understanding of the complexity of a disaster recovery.
Make a checklist, and make sure that everything is covered.
As Lync is installed or upgraded, the schema is extended with attributes. These attributes have Lync-related information in them for each Lync-enabled user or contact (or other objects). Without a proper backup of AD, there is no way to recover from a total disaster, not without starting from scratch anyway.
Active Directory is the place where new servers look for information about the location of the Central Management Store.
Think of all A-records, CNAME-records, and SRV-records of servers, gateways, pools, and simple URLs. Some of them can be autocreated when servers are added to the domain. However, a lot of them are added manually.
Make sure that you have at least a proper documentation of the internal and external DNS zone, so that the configuration can easily be recreated.
Depending on the overall backup scheme, this backup might be covered by backing up AD if the internal zones are integrated in AD.
Devices such as Lync Phone Editions use Dynamic Host Configuration Protocol (DHCP), and this is sometimes overlooked when it comes to backup. Within DHCP is the configuration for these devices on how to connect to the Lync pool. Make regular backups of the DHCP configuration as well. Call Admission Control; otherwise, E-9-1-1 might be implemented, and it would clearly give the administrator a real headache if the subnets are shifted after a recovery.
A proper backup and recovery plan of a public key infrastructure (PKI) is a complete chapter in itself. I just want to make you aware of all the certificates used in the internal deployment to create service and trusts; Lync deployments are dependent on the PKI infrastructure. With the loss of the PKI, the entire PKI will have to be recreated, and all certificates have to be recreated and deployed.
The Lync file share is covered earlier in this chapter, but the server itself also needs to be restored if disaster strikes. This can be easily done in a virtualized environment with a snapshot every now and then.
Again, the different Lync databases are already covered, but the server running SQL software should also be on the list for disaster recovery. How do we restore databases if the SQL server should fail? One way of mitigating this risk is using clustering or mirroring on the backend.
To save time in the event of a disaster, it would help to have the capability to restore the Lync Servers without installing everything from scratch. Think of the actual OS of the physical or virtual machines that run the services. If Lync is in a virtualized environment, snapshots are a great starting point.
Lync depends on Exchange in the following ways:
- Users have their mail accounts in Exchange
- Users might have their conversation history stored in their mailbox
- Their presence is synchronized from their calendar
- Contacts are stored in the Unified Contact store
- Archiving is a component that can be collocated with Exchange 2013
- The Unified Messaging service might be in use
Pay close attention to this in the recovery plan. Think of servers, roles, configurations, databases, and mail flows.
Firewalls are also an important piece of the puzzle. Some are only external, others are internal or a mix of the two. Nevertheless, we need to be able to recreate firewalls and the rules in effect for all clients and servers. The recovery plan should include backups and separate documentation of your firewalls.
The network is as important as anything else for the Lync installation. A disaster recovery plan should have a detailed description of how to back up and restore network connectivity exactly as it was, as soon as possible.
Again, think how bad it would be for elements such as QOS, CAC, and 9-1-1 if things got lost or recreated in the wrong way. Bad quality calls could start happening, or calls could erroneously be denied. A 9-1-1 call might be sending the wrong information to the emergency services and dispatch rescuers to the wrong address.
As with the firewalls, make sure that you have backups and a detailed documentation.
A reverse proxy is a required component to enable and allow features such as mobility, remote login, and federation to work properly. Remember to back up these services as well.
To complete the disaster recovery plan, have a clear idea of where to find proper installation media for Windows Server, SQL, Lync, clients, and so on. I know it can be downloaded, but we might face Internet connectivity issues at the point of recovery. Have it all downloaded and updated with the current releases and have them included in the backup.
As the recovery plan is created, try to think of other elements in the organizations infrastructure that are not mentioned earlier. There might be some kind of a third-party application or solution that the organization depends on. It should all be there on the checklist, next to all of the items described earlier.