Thank you.  Unfortunately this won't work because 0.21 is already being
creating:
~# ceph pg force_create_pg 0.21
pg 0.21 already creating


I think, and I am guessing here since I don't know internals that well,
that 0.21 started to be created but since its OSD disappear it never
finished and it keeps trying.

On Sun, Jun 7, 2015 at 12:18 AM, Alex Muntada <[email protected]> wrote:

> Marek Dohojda:
>
> One of the Stuck Inactive is 0.21 and here is the output of ceph pg map
>>
>> #ceph pg map 0.21
>> osdmap e579 pg 0.21 (0.21) -> up [] acting []
>>
>> #ceph pg dump_stuck stale
>> ok
>> pg_stat state   up      up_primary      acting  acting_primary
>> 0.22    stale+active+clean      [5,1,6] 5       [5,1,6] 5
>> 0.1f    stale+active+clean      [2,0,4] 2       [2,0,4] 2
>> <reducted for ease of reading>
>>
>> # ceph osd stat
>>      osdmap e579: 14 osds: 14 up, 14 in
>>
>> If I do
>> #ceph pg 0.21 query
>>
>> The command freezes and never returns any output.
>>
>> I suspect that the problem is that these PGs were created but the OSD
>> that they were initially created under disappeared.  So I believe that I
>> should just remove these PGs, but honestly I don’t see how.
>>
>> Does anybody have any ideas as to what to do next?
>>
>
> ceph pg force_create_pg 0.21
>
> We've been playing last week with this same scenario: we stopped on
> purpose the 3 OSD with the replicas of one PG to find out how it affected
> to the cluster and we ended up with a stale PG and 400 requests blocked for
> a long time. After trying several commands to get the cluster back the one
> that made the difference was force_create_pg and later moving the OSD with
> blocked requests out of the cluster.
>
> Hope that helps,
> Alex
>
_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Reply via email to