rpc.svcgssd segfaults on client connect. ltrace looks like this: [...] gss_krb5_set_allowable_enctypes(0x7fff85b07820, 0x55e19f865e20, 7, 0x55e19f869650) = 0 gss_accept_sec_context(0x7fff85b07a90, 0x7fff85b07aa0, 0x55e19f865e20, 0x7fff85b07ad0) = 0 gss_display_name(0x7fff85b07a94, 0x55e19f874680, 0x7fff85b07b40, 0x7fff85b07ac8) = 0 calloc(41, 1) = 0x55e19f877700 memcpy(0x55e19f877700, "nfs/lorbeer.horstn.dynv6.net@PHU"..., 40) = 0x55e19f877700 __vfprintf_chk(0x7fd7105464e0, 1, 0x55e19dd7ffed, 0x7fff85b07930sname = nfs/lorbeer.horstn.dynv6.net@PHUK.ATH.CX ) = 49 gss_release_buffer(0x7fff85b07a94, 0x7fff85b07b40, 49, 0x7fd710547710) = 0 memcmp(0x7fd7107c46de, 0x55e19dd7fbff, 9, 0x55e19f864010) = 0 nfs4_gss_princ_to_ids(0x55e19df83030, 0x55e19f877700, 0x7fff85b07a98, 0x7fff85b07a9c <no return ...> --- SIGSEGV (Segmentation fault) --- +++ killed by SIGSEGV +++ Which looks a lot like this issue: https://patchwork.kernel.org/patch/10230967/ The exact patch wasn't applied, but the bug was fixed later in this commit: http://git.linux-nfs.org/?p=steved/nfs-utils.git;a=commit;h=f3d67894cb8d785d4c12d0fa708bdddad0372de5 Upgrading to 2.3.3 solved the issue for me, but 2.3.2 should work as well. Maybe consider for stabilization, because to me it looks like the current stable version is broken with kerberized NFSv4.
I can confirm the breakage. Applying the upstream patch to nfs-utils-2.3.1-r3 fixes it for me.
