So i have been working on this case for about 4 weeks. Where a client have been having issues with CSV’s becoming unhealthy and Cache drives going into Lost Communication during backup. We have a MS Support case but as the client do not have Premier Support it’s taking a while to get trough 1st line and getting it escalated.
The problem started after a patch cycle of there Storage Spaces Direct cluster which has ReFS CSV volumes. This week i was going trough the specifications for Backup Exec and noticed this little line.
1. Veritas does not support Cluster Shared Volumes that are created on ReFS volumes or Cluster Shared Volumes with data deduplication enabled in Windows
Now this is a pretty crucial line and informed the client, wich at the time had removed all CSV backups except for 2 linux machines and gone over to agent backups. Today another person posted a thread on storagespacesdirect.slack.com saying that they where having issues during backup where drives would go into Lost Communication, they where also using Backup Exec. They will also go away from CSV backup and over to agent based backup to test this out. During the weekend i will have a result from this. But it looks like something happend in march/april/may patches that is making Backup Exec create problems for a ReFS CSV volume. And causing drives to go into Lost Communication.
So if you are using Backup Exec and have Cluster Shared Volumes with ReFS and are on 2016 but not installed latest patches you might get into trouble soon.
To fix the Lost Communication issue. You will need to reboot the node with the Lost Communication status. But wait for any repair job to finish before pausing and rebooting the node.