TrialandError Posted July 13, 2015 Report post Posted July 13, 2015 We are running SCCM 2012 R2 with 1 CAS and 3 DP's. Starting approximately 3 weeks ago I started getting reports of computers failing OSD during driver provisioning with error code 0x80072ee2 (operation timed out). I cannot think of anything that changed during that time. The problems happens with any TS and only at the driver provision step. We are using auto driver provisioning and use categories to keep them organized. Here is what I have come up with so far: - I am able to ping the DP's during the TS failure. - The clients are able to see the DNS and DHCP servers. - I turned off antivirus and the firewall on the distribution point temporarily but this had no effect. - This is happening on several different models all of which have the proper drivers imported. - If I use driver packages then the TS finishes successfully. - All of our SCCM servers are virtual so I tried migrating DP2 to a different host but this had no change on the behavior. - I changed the Network Access account but saw no improvement. - I added the 2 variables SMSTSDownloadRetryCount and SMSTSDownloadRetryDelay as a work around so that the client will keep trying to get the drivers and this has helped but the underlying issue is still there. - Comparing the client SMSTS log and the MP_DriverMgr.log shows that the client sends a driver request and it is received by the DP. The DP executes a driver match but doesn't send a response back to the client for ~8 minutes. By that time the client has already failed the driver step with the Timed Out error. I have attached logs from the client and the DP. - This is only happening at Distribution Point 2. As a test, I changed our main boundary group to assign clients to Distribution Point 1 and imaged a couple computers. The computers looked to DP1 for drivers and DP executes the driver request and sends the response in under 2 minutes. I know that this confusing fast so if you have any questions just ask and I'll get more specific where needed. I have edited the hostnames out of the logs. SMSTSExample.log DP_DriverMGR.log Quote Share this post Link to post Share on other sites More sharing options...
TrialandError Posted July 16, 2015 Report post Posted July 16, 2015 Well I think I have figured this out. It looks like the maintenance task 'Rebuild Indexes' was not actually doing anything. I disabled the task in SCCM and created a Maintenance Plan to reorganize, rebuild, and update the indexes on Site 2's database. I scheduled the task to run during off hours and made sure there were no other task scheduled to run at the same time. I also took a full DB back up and spoke with our backup engineer and asked him to run a full back up on the server prior to my maint plan kicking off. The maintenance plan ran and successfully rebuilt the indexes. 10 computers have been imaged since then and the response time on driver request is back down to below 2 minutes. I am still going to monitor the MP_drivermgr log but right now everything is working as it should. I am assuming that the DB was so fragmented that it was just taking a long time to parse for drivers. Here are some links that I read over in preparation of creating my maint plan. http://blogs.technet.com/b/smartinez/archive/2014/03/28/talking-database-in-configmgr.aspx https://stevethompsonmvp.wordpress.com/2014/05/19/powershell-sql-audit-script/ https://ola.hallengren.com/sql-server-index-and-statistics-maintenance.html Quote Share this post Link to post Share on other sites More sharing options...
