I'm still argumenting back and forth, what is the best casesensitivity mode for a dataset, that is used primarily for sharing files via APF to Mac clients. As Apple is slowly moving towards SMB this might change to SMB sometimes in the future.
As HFS+ is case insensitive, casesensitivity=insensitive sounds to be the first choice. But as I've run my FreeNAS server with casesensitivity=sensitive for quite some time, there may already be files with mixed case sensitivity on the server. So casesensitivity=mixed might be better?
But is there really any important benefit in real world usage for switching from sensitive to mixed, which is quite a bit of a hassle to do. I've reread the zfs documentation about case sensitivity several times, but I still don't get, how the system behaves with AFP requests to a case sensitive vs mixed zfs filesystem. IMO if I write a file foo to the share while a file FOO exists already, with insensitive it overwrites FOO with sensitive it creates a new file and with mixed? And if foo and FOO exist both, what does a mixed filesystem show me? Two files or one, and if only one which one?
It seems, sensitive probably gives more predictical results. As long as the files stay on the server, there are no problems with case sensitivity set to sensitive. Only, when mixed case files are transferred to the HFS+ volumes, conflicts occur. But the Finder warns you about the issue. So currently I tend to stick with sensitive.
Would by nice, if someone could clarify the behavior with casesensitivity set to mixed and give some recommendations, which mode to use on the zfs side for my kind of usage.
Pete
As HFS+ is case insensitive, casesensitivity=insensitive sounds to be the first choice. But as I've run my FreeNAS server with casesensitivity=sensitive for quite some time, there may already be files with mixed case sensitivity on the server. So casesensitivity=mixed might be better?
But is there really any important benefit in real world usage for switching from sensitive to mixed, which is quite a bit of a hassle to do. I've reread the zfs documentation about case sensitivity several times, but I still don't get, how the system behaves with AFP requests to a case sensitive vs mixed zfs filesystem. IMO if I write a file foo to the share while a file FOO exists already, with insensitive it overwrites FOO with sensitive it creates a new file and with mixed? And if foo and FOO exist both, what does a mixed filesystem show me? Two files or one, and if only one which one?
It seems, sensitive probably gives more predictical results. As long as the files stay on the server, there are no problems with case sensitivity set to sensitive. Only, when mixed case files are transferred to the HFS+ volumes, conflicts occur. But the Finder warns you about the issue. So currently I tend to stick with sensitive.
Would by nice, if someone could clarify the behavior with casesensitivity set to mixed and give some recommendations, which mode to use on the zfs side for my kind of usage.
Pete
Last edited: