Hi Sage,

i uploaded the query to http://yadi.sk/d/XoyLElnCDrc6Q

last time, after i saw "slow request" in osd.4,
i removed, format and re-added osd.4

but after i saw this query, i found many "acting 4". and i think that
indicate that this PG was in osd.4 before.

currently pg 4.7d acting in 0,6 but its empty
now the state changed into down+incomplete

have any idea ?

btw how can i map rbd volume (openstack) into physical node,
i tried using rbd map but show error no such devices ?
but there is no error if i create volume directly using rbd.

i'm just thinking whether i can take out the data inside the volume ?

Thanks
Rzk



On Thu, Dec 5, 2013 at 11:36 PM, Sage Weil <[email protected]> wrote:

> ceph pg 4.7d query
>
> will tell you which OSDs it wants to talk to in order to make the PG
> complete (or what other information it needs).
>
> sag
>
> On Thu, 5 Dec 2013, Rzk wrote:
>
> > Hi All,
> >
> > I found 6 pgs incomplete while "ceph health detail" after 3 osds down,
> > but after i manage to start again all 3 osds, only 1 left incomplete pg.
> >
> > root@:~# ceph health detail | grep 4.7d
> > pg 4.7d is stuck inactive for 306404.577611, current state incomplete,
> last
> > acting [6,0]
> > pg 4.7d is stuck unclean for 306404.577947, current state incomplete,
> last
> > acting [6,0]
> > pg 4.7d is incomplete, acting [6,0]
> >
> > does anyone know how to fix/repair this "incomplete" state ?
> > i have tried to add more osd, done almost all suggestion i found in
> google,
> > but no luck.
> >
> > and also i can't start any instance in the openstack.
> >
> > Thanks,
> > RzK
> >
> >
>
_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Reply via email to