We are experiencing issues with server access that seems to be related to Shadow Copy. This is being seen multiple servers with anywhere from 20 to 35TB of storage at least 10TB of data on the data drive. I have been able to reproduce what I am seeing on Server 2012, Server 2016 and Server 2008.
Here is a description of what I am seeing. We have a Shadow Copy run at noon. A little after 12:05 we see access to the server appear to freeze. This goes on for about 3 to 5 minutes. During this time copy jobs, processes being run on the server and even trying to delete small files will halt/pause. Sometimes processes will crash, and we are seeing copy jobs fail as well. We will often run processes that take days to complete and copy jobs are often of multiple terabytes. It looks to me that the actual Shadow Copy job has completed when we see the freezing. That is my assumption because when I look at the Shadow Copy job in Computer Management it shows a time of 10:05 for the current job which I am assuming is the time the Shadow Copy itself completed.
In reading about Shadow Copy it does say that during the process it will freeze files in order ready the file system, but if I understand correctly that should happen at the beginning of the process and should only last few seconds and should not be allowed to last longer the 60 seconds.
During the Shadow Copy process the write I/O request remain frozen but should the process should only last 10 seconds
So I am wondering if there is an issue with the way Shadow copy is completing its task or if maybe this is to be expected due to the quantity of data on the servers.
