I just upgraded our seafile cluster from pro 7.0.19 to pro 9.0.7 on Centos9 with an NFS backend. I enabled the golang fileserver, and now I see lots of “internal server errors” in the fileserver log, like
path /repo/9b5c9e24-.../commit/HEAD/ internal server error: Failed to get commit 2dd7f... for repo
path /repo/eded9... internal server error: Failed to close block eded...
Shall I go back to the normal fileserver?