Re: Uncompactable Monitor Store at 69GB -- Re: Cluster in warn state, not sure what to do next.

["wido@xxxxxxxx" <wido@xxxxxxxx>]

Op 21 jul. 2016 om 21:06 heeft Salwasser, Zac <zsalwass@xxxxxxxxxx> het volgende geschreven:

Thanks for the response!  Long story short, there’s one specific osd in my cluster that is responsible, according to the dump command, for the two pg’s that are still down.

 

I wiped the osd data directory and recreated that osd a couple of days ago, but it is still stuck in the “booting” state.  Any ideas how I can investigate that particular osd further?

 

 

Did you re-use the old UUID of the OSD as set in the OSDMap? Otherwise it will stay in booting.

However, wiping an OSD like that is usually not a good idea.

Wido

 

From: Gregory Farnum <gfarnum@xxxxxxxxxx>
Date: Thursday, July 21, 2016 at 3:01 PM
To: "Salwasser, Zac" <zsalwass@xxxxxxxxxx>
Cc: "ceph-users@xxxxxxxxxxxxxx" <ceph-users@xxxxxxxxxxxxxx>, "Heller, Chris" <cheller@xxxxxxxxxx>
Subject: Re: [ceph-users] Uncompactable Monitor Store at 69GB -- Re: Cluster in warn state, not sure what to do next.

 

On Thu, Jul 21, 2016 at 11:54 AM, Salwasser, Zac <zsalwass@xxxxxxxxxx> wrote:

Rephrasing for brevity – I have a monitor store that is 69GB and won’t

compact any further on restart or with ‘tell compact’.  Has anyone dealt

with this before?

 

The monitor can't trim OSD maps over a period where PGs are unclean;

you'll likely find that's where all the space has gone. You need to

resolve your down PGs.

-Greg

 

_______________________________________________
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