On 1/12/22 11:13, Jeremy Allison wrote:> On Wed, Jan 12, 2022 at 03:07:05PM +0100, Ralph Boehme via samba wrote: >> >> no, it should work afaict. Just beware that afair ZFS on Linux doesn't >> support NFSv4 ACLs, just POSIX ACLs. As Samba will work with POSIX >> ACLs by default, other ACL flavours require loading a dedicated ZFS >> module, your setup should basically just work. But ff course, the >> devil's in the details, so you should do some decent researcg abd >> testing before deploying a production system. :) > > If you're planning this from scratch as a Samba-only filesystem, > I'd really recommend setting the filesystem up as case insensitive. > You'll get much better performance on large directories that way.Since we do a lot of computational science, almost nothing is Samba only, and I most certainly can't assume this even in cases where it appears to be Samba only. One or more linux boxes with NFS mounts always show up eventually. <:) I'm curious about why there are performance benefits to ignoring case, but I suppose it's because file inquiries involve case mapping to accommodate Windows machines. In the age of Unicode, ignoring case is an awful, awful idea, but Microsoft is now stuck with this legacy implementation and I don't see how they'll ever escape it. Maybe once Windows is rebased on a linux kernel... Thanks to everyone who chimed in on this; my use case is POSIX ACLs (it's the greatest common denominator), so looks like this will all be fine. @Roland: Linux and BSD now share a common code base (zfsonlinux is now OpenZFS), so your comment would indicate that vfs_zfsacl is dead, since Solaris appears to be dead.>>> This message is from an external sender. Learn more about why this << >>> matters at https://links.utexas.edu/rtyclf.??????????????????????? << >
On Wed, Jan 12, 2022 at 02:33:45PM -0600, Patrick Goetz via samba wrote:> >I'm curious about why there are performance benefits to ignoring case, >but I suppose it's because file inquiries involve case mapping to >accommodate Windows machines. In the age of Unicode, ignoring case is >an awful, awful idea, but Microsoft is now stuck with this legacy >implementation and I don't see how they'll ever escape it. Maybe once >Windows is rebased on a linux kernel...The reason it is a performance win is in the lookup of an unknown file case. If a client requests to open "Filename", and a stat() of "Filename" fails, then smbd has to open and scan the directory looking for any file named "filename" as a case-insensitive match. For a large directory that can be very time-intensive. MacOSX is also case-insensitive I believe :-(.
On 1/12/22 21:33, Patrick Goetz via samba wrote: > @Roland: Linux and BSD now share a common code base (zfsonlinux is now > OpenZFS), so your comment would indicate that vfs_zfsacl is dead, since > Solaris appears to be dead. Not really. First, OpenZFS is in FreeBSD 13, with 12 still supported until mid 2024. Second, FreeBSD still has NFSv4 ACLs, even after the switch to OpenZFS. For the man page:> acltype=off|nfsv4|posix > Controls whether ACLs are enabled and if so what type of ACL to use. > When this property is set to a type of ACL not supported by the current > platform, the behavior is the same as if it were set to off. > off default on Linux, when a file system has the acltype > property set to off then ACLs are disabled. > noacl an alias for off > nfsv4 default on FreeBSD, indicates that NFSv4-style ZFS ACLs > should be used. These ACLs can be managed with the > getfacl(1) and setfacl(1). The nfsv4 ZFS ACL type is not > yet supported on Linux. > posix indicates POSIX ACLs should be used. POSIX ACLs are > specific to Linux and are not functional on other > platforms. POSIX ACLs are stored as an extended > attribute and therefore will not overwrite any existing > NFSv4 ACLs which may be set. > posixacl an alias for posixbye av.