Skip to main content

What to do when your application server goes bang

What happens when someone kills your Application Server?

So imagine the scenario:

Three Server Solution
- SQL
- SharePoint 2010 and Project Server 2010 Application Server (Central Admin Host)
- Wfe/ReportServer

We wake one bleary Monday morning to find that some numpty has killed the application server and the users are baying for blood.

Well surprisingly SharePoint handles this disaster recovery scenario particularly well.  Well.  Better than I thought it would to be honest.

Rough steps:

- quick SQL backup to be safe
- Rebuild your application server
- reinstall pre-reqs
- reinstall SP, PS, SPFSP1, SPS+PSSP1, Cumulative Update and other stuff you usually put on there.
- Run your configuration wizard to reattach to the Farm, and select Host CA Site

The last step was what I was VERY wary of.  Would the server simply reattach to the Farm, even when there is no CA server available?  

Bingo your back.... almost.... you are going to get errors a-gogo in your event log as things just aren't quite back to normal after the farm had its head cut off.

Error:  Navigate to your Project Server site and select BI Center:  Unexpected error occurred
Logs:  URL events stating System.ServiceModel.ServerTooBusyException:  Secure Store Service Did Not Performed (sic) the operation
Fix:       Secure Store needs to be reset 

> Service Applications | Secure Store Properties
> Regenerate Key

Just for safety i went through and reset all properties for configured Service Applications.

Started Microsoft SharePoint Foundation Web Application Service as IIS was also not showing the same sites on application server as on the WFE.

Outstanding issues:
- Navigating to performancepoint content returns an Unexpected Error & Event Log shows
"the user xxxxx attempted to access an item in the following location http://..............."
Verify the location exists and that the user has the Read Items permission

Ok so I missed resetting the PerformancePoint Service Application properties... after doing this, all is well

Search:
Application SErver Administration Job Failed for SErvice instance Microsoft.office.server.search.administration.searchServiceInstance
Reason: There is no project Portal_Content mounted under gatherer application 68d05007-b712-46f2-adc6-58c22066dbd0.

I didnt find a way around this... recreated Search Service Application

Anyway that was my route to getting this farm up and running again.  we were lucky we didnt have any code recover from the dead app server.


At the end of this process we had:
- SP Sites back up and available
- CA Back up and available
- PWA sites and BI reports


Hope it helped some...


Carl

Comments

Popular posts from this blog

Issues update on #projectserver2013 - Timesheets and Publishing

Reporting Publish ** updated with links to other related discussions, and a VBA macro ** the following issue has been noted on publish since June 13 CU was applied: ReportingProjectChangeMessageFailed (24006) - Object reference not set to an instance of an object.. Details: id='24006' name='ReportingProjectChangeMessageFailed' uid='4d869e56-f625-e311-bb41-005056b90052' QueueMessageBody='Project UID='e3f49977-b2bc-e211-8559-005056b90052'. PublishType='ProjectPublish'' Error='Object reference not set to an instance of an object.'. I have seen this noted previously on a similar issue: http://nearbaseline.com/blog/2013/06/ms-reporting-project-publish-jobs-failed-after-aprilcu/comment-page-1/#comment-14741 This is  caused by Baselined Milestones having NULL Baseline Cost values Original bug note with potential workaround is here: http://blogs.msdn.com/b/brismith/archive/2012/05/23/project-server-2007-reporting-project-pu...

#projectserver2013 VIEW FAILURE: The view failed to load. Press OK to reload this view with the default settings. Press cancel to select another view.

** UPDATE ** includes notes relating to secondary bug where Timesheet is created without Administrative tasks. Does this ring any bells? This has been bugging me for months, but finally I have a repro for this: Issue Summary:  When a task is deleted from a plan that is approved into a previous or current timesheet - even when there are no actuals on the task - you can no longer view the timesheet The following repro has been proven: - Setup system with Single Entry Mode, with enforced Status Approval before Timesheet Approval - Create resource as own timesheet manager - Create new project - Create two tasks in the same week, starting monday with 5 days duration:  1) Task to assign actuals, 2) Task to delete post submission - Assign Resource to tasks - publish project - as Timesheet User, go to the appropriate timesheet period for the tasks created - Assign actual work to one task (task 1), leaving task 2 with no actual work - Submit timesheet - as Project ...

Restoring PWA Site to another Web App in the same Farm

The scenario is this: SharePoint 2016 Farm with Project Server Two Web Apps Development UAT One PWA on Development Web App. I want to copy the PWA Site on Development web app to UAT to support a testing cycle. As far as I knew there were two options: 1) Content Database Restore and Attach Process would be backup your Dev Content Database, Restore to a new Content Database for QA, then mount the database on the appropriate web app and your off.... Problem:  Although you can do this with the -AssignNewDatabaseID switch in Powershell (to avoid two content db's having the same database id) the Site Collection (PWA) in the db still retains its SiteID which means there is a duplicate SiteID in the Configuration Database.  This stops the PWA site being created and alllocated correctly and becomes essentially orphaned. This method is only any good for MOVING not COPYING Back to the drawing board... 2) Backup-SPSite / Restore-SPSite I didn't believe this ...