We help IT Professionals succeed at work.

disaster recovery plans for virtual systems

pma111
pma111 used Ask the Experts™
on
1) Do you back up ESXi hosts, or do you just backup the vitual machines running on the hosts?

2) In terms of your corporate diaster recovery plan document, do disaster recovery documented plans typically include technical details per system type, i.e. what do your disaster recovery plans detail for esxi hosts, vcenter etc. To what level?

3) Albeit not on a technical level, I am told often organisations have a 2 site replicated platform, so if one site goes down, the other kicks in and systems wont go down. Do you still need disaster recovery plans in that kind of setup?
Comment
Watch Question

Do more with

Expert Office
EXPERT OFFICE® is a registered trademark of EXPERTS EXCHANGE®
Top Expert 2010
Commented:
1. No to hosts because it's prob quicker to just rebuild them from scratch. But for me, that holds true due to not having many hosts. If you have...say...more than 5 hosts, you can install using Auto Deploy. So, you just need Auto Deploy procedures for your org.
1. a. Yes, I back up VMs mostly. I use Veeam B&R. I replicate VMs offsite (more on this in Q3)

2. DR Plan starts with Biz Continuity plan. DR is a subsection of BC. IT/IS needs involved in drawing up the DR plan for sure though. It doesn't necessarily to be too intricately detailed, but detailed enough to know what to do in the event of a disaster. Step-by-step procedures for DR processes should be drawn up within the IT Dept tho. But, the org DR Plan shouldn't show such details.

3. Yes, typically orgs have, and should have, offsite location for DR. We do. I use Veeam to replicate my most critical VMs to that site. All I need to do is power them up and I'm good to go. Yes, you still need a DR Plan because it may not all be seamless if the main site goes down. There are typically 'procedures' that need to be done to be fully up & operational in a DR scenario.


Regards.
~coolsport00
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
Fellow 2018
Expert of the Year 2017
Commented:
1. Yes, we Backup the flash drives using my article

How to Backup an ESXi installation on an USB Flash Drive or SD card, for security or redundancy.

we also backup the configuration of the ESXi hosts daily, using vCLI, backup commands.

2. All our DR plans can be followed by a non-IT person, they are that simple. Reason being just in case the IT department is bombed, how would anyone know how to enact DR.

3. Correct, it's often part of DR, two Sites, Site A and Site B

Author

Commented:
Does 3) manage itself, i.e. do you need a "documented plan" if site A goes down, as its already configured for site B to take charge and run the show?

2) For example vcenter and esxi hosts, what would your plan "show" in terms of detail? And how important is a document that shows what happens in the event of a disaster?
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
Fellow 2018
Expert of the Year 2017

Commented:
2) It's shows every single step, by step, the DR documents is a BOOK, which runs to about 1000 pages.

3) All steps are documented, involving if Site B is also unavailable to move to Mobile Site C, when Contractors are used Sunguard to bring in Mobile DataCentre and Generators on a 40 tonne lorry.
Top Expert 2010

Commented:
1. No, 3 doesn't manage itself. As I stated above, you need a high-level org plan, then a more detailed intra-departmental plan for the tasts/steps needed to actually fully "fail over".

2. This can only be answered by your IT team based off their infrastructure and failover plans. You should have a doc to GUIDE what needs done to failover to avoid confusion and make sure the most critical systems get online the quickest. With all the craziness going on during a disaster, having a doc is a necessity to assist in the failover process....

Author

Commented:
Ok, but in terms of laymans DR, is

2) a "server by server" type step, i.e. when recovering each server are there seperate steps involved, or are you recovering a site. I.e. does each server in your network need "step by step" recovery steps, or is it not usually that granular. Or if I asked you for your DR plan, and you manage 200 servers, will I find 200 step by step recovery stages, 1 per server?
Andrew Hancock (VMware vExpert / EE Fellow)VMware and Virtualization Consultant
Fellow 2018
Expert of the Year 2017

Commented:
The DR plan produced is to return the business to a minimal "working state". as defined by the business. The Businesses List which Services are Critical ro the running of their business, so they may not have ALL servers restored, after 5 mins.

e.g. a Client has defined the following to be crtical servers

1. ERP system - this requires restoration of SQL, Oracle Servers
2. Email - this requries restoration of Exchange, Active Directory Servers
3. Service for 200 Desktops - this requies restoration of Citrix Terminal Servers

Procedures detail how to restore a service, which may include the restortation of sefveral servers, which are highlighted with step by step procedures on how to restore each server.

Correct, there is a Chapter for Each Service, each Service is broken down into, which servers to restore and how, step by step, and in which order.