Hopefully, no one will ever need to read this, but I had a weird issue today. I’ve never had this before with Configuration Manager or any other app for that. I use a lab based on virtual machines and at the end of each day the VMs are put into a saved state as the host machine is powered down. Not usually an issue, I’ve been using a similar approach for years on laptops whilst on the move etc. Today however, I fired up my lab prior to a customer call in which I would require it and mid-call the ConfigMgr console decided to crash and not want to reconnect.
No big deal I hear you say, this kind of thing happens from time to time. I completed the call and made my apologies then later on decided to see what was going on.
- All ConfigMgr services are running
- All SQL services are running
- I can connect to SQL and run some queries against it
- I can connect to WMI on the site server and enumerate classes
Bit odd I thought. Before jumping into the ConfigMgr logs I decided to check the event viewers application log to see if there was anything happening that may break ConfigMgr. That’s where it started to get interesting as there was a sea of red.
So, lets do as instructed in that event and go see the statesys.log to see what’s going on – more red.
We can see ConfigMgr isn’t happy with SQL for some reason with “Failed to connect to the SQL Server, connection type: SMS ACCESS”
If we read through the lines I can also see that “SQL Server Network Interfaces: The clocks on the client and server machines are skewed.” Bit odd I though as they are all the same. I of course doubly checked the ConfigMgr server, the SQL server and the domain controller (I have just on in lab). All times are looking good.
Next step, when did this start happening as it crashed in the middle of my call so about 1:30PM. Scrolling up in the statesys.log I can see that actually the error was occurring prior. Right back until the point I actually started these back up from their saved state prior to my call. This is where I noticed a problem. Eagle eyes at the ready…
This sucker that I’ve highlighted had todays time and date logged right in the middle of the previous running time logging. This is 09:41:42 on 12/7/2020. Logging then continued using 12/3/2020 (4 days earlier) for another second or so at which point the correct logging snapped into place again but with the errors displayed. That kind of explains the “skewed” time we’re seeing and there’s obviously a mechanism in place to detect this condition.
So, how to fix it?
Well really simply in my case the server was on a timed power off and fortunately when this was powered up again from the saved state it all came back into sync enough for the console to connect.
I did however, still note some errors reporting in Event Viewers Application Log for Windows Server Update Services which uses the same SQL server in my case. I gave the SMS_EXECUTIVE service a restart on the site server (this will temporarily take your ConfigMgr site down) and everything came back up again perfectly.
Hopefully you’ll never need this but if you do you can resolve it with a level of confidence.