On Wed, Aug 21, 2019 at 05:55:51PM +0100, Daniel P. Berrangé wrote:
> On Wed, Aug 21, 2019 at 12:47:03PM -0400, Marcus Furlong wrote:
> > On Wed, 21 Aug 2019 at 08:23, Daniel P. Berrangé <berrange(a)redhat.com>
wrote:
> > >
> > > On Tue, Aug 20, 2019 at 11:11:07AM -0400, Marcus Furlong wrote:
> > > > Resend to libvir-list in case that is more appropriate:
> > > >
> > > >
> > > > Hi,
> > > >
> > > > I get the following error when running libvirtd on MacOS as root:
> > > >
> > > > 2019-07-11 00:12:33.673+0000: 123145573953536: error :
> > > > qemuProcessQMPLaunch:8501 : internal error: Failed to start QEMU
> > > > binary /usr/local/bin/qemu-system-x86_64 for probing: libvirt:
error
> > > > : cannot set supplemental groups: Invalid argument
> > >
> > > Are you able to run 'strace' (or whatever MacOS eqiuv is) to see
> > > the values passed to setgroups when it fails ?
> >
> > I ran `dtruss -f -l -s /usr/local/sbin/libvirtd` but the setgroups
> > calls seem to be missing.
> >
> > Looking at other sources, it seems like some have special treatment of
> > setgroups on MacOS, e.g. samba:
> >
> >
https://github.com/samba-team/samba/blob/v4-11-stable/source3/smbd/sec_ct...
> >
> > Perhaps this is needed for libvirt?
>
> The capping of ngroups to NGROUPS_MAX looks like a possibe reason.
>
> Adding this debug might show us if we're exceeding it:
>
> diff --git a/src/util/virutil.c b/src/util/virutil.c
> index 89d2cf011f..effc02b898 100644
> --- a/src/util/virutil.c
> +++ b/src/util/virutil.c
> @@ -1043,6 +1043,7 @@ virSetUIDGID(uid_t uid, gid_t gid, gid_t *groups
ATTRIBUTE_UNUSED,
> }
>
> # if HAVE_SETGROUPS
> + VIR_DEBUG("setgroups %d max %d", ngroups, NGROUPS_MAX);
> if (gid != (gid_t)-1 && setgroups(ngroups, groups) < 0) {
> virReportSystemError(errno, "%s",
> _("cannot set supplemental groups"));
>
>
Yes, there's an overflow:
2019-08-21 18:25:37.943+0000: 123145413914624: debug : virSetUIDGID:1046 : setgroups 23
max 16
Related samba ticket (it also has references to the python and dovecot
issues):
https://bugzilla.samba.org/show_bug.cgi?id=8773
The quick hack is to simply truncate groups to NGROUPS_MAX.
Reading that bug, the proper fix looks quite alot more complex
Regards,
Daniel
--
|: