0

I have a client that just wants 0 scheduled downtime with 0 data loss with our SharePoint 2013 environment.

Our plans included a load balanced 3 tiered farm (2 WFE, 2 App servers, 2 SQL servers (AlwaysOn)). The initial thought was that we sever the connections between the two parts (move all services and DBs from web1, app1, sql1 to web2, app2, sql2 and remove the capability for the two halves to communicate with each other), patch one side of the environment, bring it back online while taking the other down for patching.

While that would work for OS patching, it would be a total disaster for SharePoint upgrades since one database would have new data and the other database would be in a new version, we'd either lose the data and have a "happy farm" or we'd lose the update and have an "angry farm."

My other thought was to utilize a Hot COOP style farm where we acquiesce the current production to the COOP production, update the current production then let current production take over again, I thought this would be far less dangerous than trying to split up a farm, but we still have the same database issues.

Are there any products or services that can allow this to happen?

1

There is a supported method that Sam Betts writes about with no downtime. It essentially stops a SQL Server job that does the log-file shipping.

  • Disable "LSBackup_WSS_Content" for your DR farm
  • Switch users to the DR environment in DNS
  • Patch Primary Farm
  • Switch Back Users to Primary Farm in DNS
  • Resume Log-Shipping to DR
  • Patch DR to match Primary

http://blogs.msdn.com/b/sambetts/archive/2013/11/15/patching-sharepoint-farms-with-no-downtime-high-availability-sharepoint.aspx

  • It also involves setting the COOP site into read-only mode. It's a good option, but my concern is that our client will see this as being equivalent to being down if they can't utilize the collaborative features. – Snowburnt Nov 12 '14 at 17:29
0

It’s not possible to update or patch SharePoint without occurring some amount downtime. What we can do is to minimize the downtime. Even if you follow the Sam Bettes's guideline, still their are plenty of chances of the Downtime i.e when you change DNS and it take time to propagate, if error then you have to flush DNS on server etc.

Ideally, in my experience it is nto possible, even for Office 365 says 99.99% but usually fall little low. But follow the instruction mentioned in Mike's reply and minimize it.

read this discussion. https://social.technet.microsoft.com/Forums/sharepoint/en-US/e1c5308b-f009-4c88-998e-06709cdf9c26/client-needs-99999-uptime-including-no-downtime-during-sharepoint-and-other-patches?forum=sharepointgenerallegacy

  • I've found one thing...Metalogix Replicator might be able to do it. It replicates the farms based on SharePoint events rather than the database. If I take one down it queues up the events on one farm and sends them down to the next. It can even work with separate major versions of sharepoint (2010 to 2013) (or so they say). – Snowburnt Nov 12 '14 at 19:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.