That’s usually shown when a device has an id of “0” which is usually when it’s being removed or has only just been added.
That’s a browser refresh button press by the way. It triggers a re-read of info to be displayed in the Web-UI.
Which session is this. The SMB session?
You could check the id’s of the devices in the associated problem-pool from the command line via:
sudo btrfs fi show (optional mount point here i.e. /mnt2/pool-name)
and to check the error stats on a pool from the command line you can use:
sudo btrfs dev stats /mnt2/pool-name-here
It may be you have evidence there of a drive failing or the like.
Is this a pool using btrfs “single” raid level. In which case if there is an issue the pool will not be able to resource a duplicate copy or self repair if the duplicate copy is good (passes it’s checksum).
Also do other machines / Operating Systems show the same issue when traversing this tree. Some clients don’t like ‘seeing’ too many files.
Hope that helps. I’ll be pleased when you’ve moved this system over to our “Built on openSUSE” but I fully appreciate these things take time. But there are now many much older packages in the CentOS variant than are in our supported Leap 15.3 base used in the v4+. It may be that the issue is with the older Samba there for instance.