A geli device's physical path should be the underlying provider's with "/eli" appended, if the underlying provider has a non-null physical path itself.
A commit references this bug: Author: asomers Date: Wed Feb 14 20:15:33 UTC 2018 New revision: 329273 URL: https://svnweb.freebsd.org/changeset/base/329273 Log: geli: append "/eli" to the underlying provider's physical path If the underlying provider's physical path is null, then the geli device's physical path will be, too. Otherwise, it will append "/eli". This will make geli work better with zfsd(8). PR: 224962 MFC after: 3 weeks Differential Revision: https://reviews.freebsd.org/D13979 Changes: head/sys/geom/eli/g_eli.c head/tests/sys/geom/class/eli/Makefile head/tests/sys/geom/class/eli/misc_test.sh
A commit references this bug: Author: asomers Date: Sat Mar 10 03:34:28 UTC 2018 New revision: 330733 URL: https://svnweb.freebsd.org/changeset/base/330733 Log: MFC r329273, r329275, r329277, r329284, r329344 r329273: geli: append "/eli" to the underlying provider's physical path If the underlying provider's physical path is null, then the geli device's physical path will be, too. Otherwise, it will append "/eli". This will make geli work better with zfsd(8). PR: 224962 Differential Revision: https://reviews.freebsd.org/D13979 r329275: gpart: append partition name to the underlying provider's physical path If the underlying provider's physical path is null, then the gpart device's physical path will be, too. Otherwise, it will append the partition name, such as "/p1" or "/s1/a". This will make gpart work better with zfsd(8). PR: 224965 Differential Revision: https://reviews.freebsd.org/D14010 r329277: Add mtree entry for 329275 X-MFC-With: 329275 Sponsored by: Spectra Logic Corp r329284: zfsd: Allow zfsd to work on any type of GEOM provider cddl/usr.sbin/zfsd/zfsd_event.cc Remove the check for da and ada devices. This way zfsd can work on md, geli, glabel, gstripe, etc devices. geli in particular is useful combined with ZFS. gnop is also useful for simulating drive pulls in the ZFSD test suite. Also, eliminate the DevfsEvent class entirely. Move its responsibilities into GeomEvent. We can get everything we need to know just from listening to GEOM events. lib/libdevdctl/event.cc Fix GeomEvent::DevName for CREATE events. Oddly, the relevant field is named "cdev" for CREATE events but "devname" for disk events. Relnotes: Yes (probably worth mentioning the geli part) Sponsored by: Spectra Logic Corp r329344: Optimize zfsd for the happy case If there are no damaged pools, then ignore all GEOM events. We only use them to fix damaged pools. However, still pay attention to ZFS events. X-MFC-With: 329284 Sponsored by: Spectra Logic Corp Changes: _U stable/11/ stable/11/cddl/usr.sbin/zfsd/case_file.cc stable/11/cddl/usr.sbin/zfsd/case_file.h stable/11/cddl/usr.sbin/zfsd/zfsd.cc stable/11/cddl/usr.sbin/zfsd/zfsd_event.cc stable/11/cddl/usr.sbin/zfsd/zfsd_event.h stable/11/etc/mtree/BSD.tests.dist stable/11/lib/libdevdctl/event.cc stable/11/sys/geom/eli/g_eli.c stable/11/sys/geom/part/g_part.c stable/11/tests/sys/geom/class/Makefile stable/11/tests/sys/geom/class/eli/Makefile stable/11/tests/sys/geom/class/eli/misc_test.sh stable/11/tests/sys/geom/class/part/
Little point to MFCing to stable/10, since stable/10 lacks zfsd.