Ahh... I've found how to fix this up.
With ceph pg dump I can see which individual PGs have bogus statistics. Then by manually triggering a shallow scrub on the affected PGs the stats are corrected to 0 objects.
So the 8EB issue is fixed, but it's still a mystery how the PGs got into this state.
Cheers, Dan
On 28 Oct 2016 10:08, "Dan van der Ster" <dan@xxxxxxxxxxxxxx> wrote:
>
> Hi Goncalo,
>
> Strange, now ceph df says the pool has 0 bytes used and -1 objects. rados df agrees this those numbers.
>
> Cheers, Dan
>
> On 28 Oct 2016 00:47, "Goncalo Borges" <goncalo.borges@xxxxxxxxxxxxx> wrote:
> >
> > Hi Dan...
> > Have you tried 'rados df' to see if it agrees with 'ceph df' ?
> > Cheers
> > G.
> >
> > ________________________________________
> > From: ceph-users [ceph-users-bounces@xxxxxxxxxxxxxx] on behalf of Dan van der Ster [dan@xxxxxxxxxxxxxx]
> > Sent: 28 October 2016 03:01
> > To: ceph-users
> > Subject: ceph df show 8E pool
> >
> > Hi all,
> >
> > One of our 10.2.3 clusters has a pool with bogus statistics. The pool
> > is empty, but it shows 8E of data used and 2^63-1 objects.
> >
> > POOLS:
> > NAME ID USED %USED MAX AVAIL OBJECTS
> > test 19 8E 0 1362T 9223372036854775807
> >
> > Strangely, ceph df -f json doesn't agree on this value:
> >
> > # ceph df -f json | jq .pools[0].stats.objects
> > 9223372036854776000
> >
> > Even if I write some data there, the values remain bogus.
> >
> > I created a new empty pool, and it looks ok:
> >
> > NAME ID USED %USED MAX AVAIL OBJECTS
> > test2 66 0 0 1362T 0
> >
> > Anyone have any ideas what could be wrong? or how to check the pool
> > stats in more detail?
> >
> > Cheers, Dan
> > _______________________________________________
> > ceph-users mailing list
> > ceph-users@xxxxxxxxxxxxxx
> > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
