Hello  llya and paul,

Thanks for your reply. Yes, you are right, 0x7fddff8ee8cbffb is come from
kernel upgrade, it's reported by a docker container
(digitalocean/ceph_exporter) use for ceph monitoring.

Now upmap mode is enabled, client features:

"client": {
        "group": {
            "features": "0x7010fb86aa42ada",
            "release": "jewel",
            "num": 6
        },
        "group": {
            "features": "0x3ffddff8eea4fffb",
            "release": "luminous",
            "num": 6
        }
    }

Thanks again.

Ilya Dryomov <idryo...@gmail.com> 于2019年10月21日周一 下午6:38写道:

> On Sat, Oct 19, 2019 at 2:00 PM Lei Liu <liul.st...@gmail.com> wrote:
> >
> > Hello llya,
> >
> > After updated client kernel version to 3.10.0-862 , ceph features shows:
> >
> > "client": {
> >         "group": {
> >             "features": "0x7010fb86aa42ada",
> >             "release": "jewel",
> >             "num": 5
> >         },
> >         "group": {
> >             "features": "0x7fddff8ee8cbffb",
> >             "release": "jewel",
> >             "num": 1
> >         },
> >         "group": {
> >             "features": "0x3ffddff8eea4fffb",
> >             "release": "luminous",
> >             "num": 6
> >         },
> >         "group": {
> >             "features": "0x3ffddff8eeacfffb",
> >             "release": "luminous",
> >             "num": 1
> >         }
> >     }
> >
> > both 0x7fddff8ee8cbffb and 0x7010fb86aa42ada are reported by new kernel
> client.
> >
> > Is it now possible to force set-require-min-compat-client to be
> luminous, if not how to fix it?
>
> No, you haven't upgraded the one with features 0x7fddff8ee8cbffb (or
> rather it looks like you have upgraded it from 0x7fddff8ee84bffb, but
> to a version that is still too old).
>
> What exactly did you do on that machine?  That change doesn't look like
> it came from a kernel upgrade.  What is the output of "uname -a" there?
>
> Thanks,
>
>                 Ilya
>
_______________________________________________
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Reply via email to