2. Greg Kamer – The Mirazon Group MCSE:2003, MCTS: MOSS 2007, WSS 3.0 5 Years in IT 3 with SharePoint Twitter: @gregkamer Email: greg.kamer@mirazon.com About Me
3. High Availability Simply, keeping a system up as long as possible without interruption. Redundancy is key here. Tends to be more costly, but when weighed with mission critical apps, the costs may be easily justified. Some Definitions
4. Disaster Recovery How can we recover, in terms of money, speed and manpower, a system that has been brought down. Factors can be as small as someone tripping over a network cable and yanking it out to a large scale natural disaster. Costs range from small to VERY high. Definitions (cont’d)
5. Considerations must be made for the roles at hand (WFE, Excel, Indexer, Query, etc) Clustering v Load Balancing No roles in SharePoint can be clustered, with the exception of the database “role” Some roles will self-load balance Discussion High Availability in SP
6. Web Front Ends are special They must be load balanced to make them highly available This is where SharePoint and the user(s) directly meet http://geekswithblogs.net/emanish/archive/2007/05/22/112682.aspx WFEs and Load Balancing
7. Network Load Balancing – Microsoft Free, but limited in functionality Inclusive walkthrough - http://tinyurl.com/mgs566 Virtual Appliances Some free, some paid. Your mileage may vary. Hardware F5, CoyotePoint, Cisco Awesome features but pricey Load Balancing WFEs
8. When this is setup in SharePoint, you MUST specify a conversion load balancer. There are no configuration settings once the service is in place. Document Conversion
9. Multiple ECS Servers can be used. There is configuration and options, but they are few. Excel Calculation Services
10. Once a workbook is opened, the book uses the same server throughout the duration. For further study: http://blogs.msdn.com/excel/archive/2008/12/12/excel-services-and-load-balancing.aspx ECS Cont’d
11. This role cannot be made highly available. There are ways to make it somewhat highly available, but it is not an automatic process and can tax resources. Index
12. The role can run on a WFE or on separate servers The roles are made redundant by using multiple WFEs If not on dedicated WFEs, then round robin is used to serve queries Beware the dreaded “My query role runs on the index server, but he has no friends cause he doesn’t know how to share” scenario. Query Role
13. SQL Clustering – very highly available, but requires shared storage (SAN) SQL Mirroring – not automatic failover for SharePoint, nor really high availability, but can be scripted for more-highly-available-than-not scenario Database
14. Can provide instant HA in some circumstances Most any role can be virtualized *with exceptions of course* DR can be factored in here as well. Virtualization
15. We all have a plan and we have all tested it, right? Right??? SharePoint has many facets with respect to DR. SQL, Server 2008, hardware and connectivity, etc Disaster Recovery
16. How do we look at SharePoint in terms of DR? SQL? Virtualization? Redundancy? Network Connectivity? Backup and restore procedures? Hot/warm/cold recovery sites? Brainstorm
17. CRUCIAL to any DR discussion. If you have no backups, you have no disaster recoverability. Restore time and SLAs will determine the type of backups that may be required. Backups
18. Are available via the GUI (Operations -> Backup) Are really just database dumps Important for Enterprise search-gets index file and DB at the same time Not able to be scheduled! STSADM –o backup Able to be made a scheduled task Can cause table locking issues on DB Not good for large site collections Third party (AvePoint, Idera, DPM, etc) Backups
19. SharePoint native backups not able to restore individual sites, lists, or items (without a lot of hassle!) Third party tools must be used for that SPD can backup sites…but not the cleanest way to do it. Recycle Bin – Can save your bacon! Backups Cont’d
20. Creating a mirror farm. Database mirroring or log shipping Backup and restores Third party mirroring tools AvePointDocAveHigh Availability - http://www.avepoint.com/sharepoint-high-availability-docave/ Other DR considerations
21. Plan for Redundancy – MSFT Whitepaper http://technet.microsoft.com/en-us/library/cc263044.aspx Case Study of High Availability for SharePoint Server by Using Virtualized Environments and Database Mirroring – MSFT Whitepaper http://technet.microsoft.com/en-us/library/ee335343.aspx SharePoint Disaster Recovery Guide By John Ferringer and Sean McDonough http://www.amazon.com/SharePoint-2007-Disaster-Recovery-Guide/dp/1584505990/ref=sr_1_1?ie=UTF8&s=books&qid=1250795045&sr=8-1 Further Reading
22. SharePoint Backup and Disaster Recovery – Joel OlesonTechEd 2008 PPT http://www.slideshare.net/joeloleson/sharepoint-backup-and-disaster-recovery-presentation Differences between Log Shipping and Mirroring http://blogs.msdn.com/mikewat/archive/2007/07/28/database-mirroring-and-log-shipping-which-is-better.aspx More Reading
23. Q&A Please fill out your evaluations Don’t be afraid to play, it’s a resilient platform. Thank you!