Re: OSD down after PG increase

Dan Van Der Ster <daniel.vanderster@xxxxxxx> · Thu, 13 Mar 2014 10:19:06 +0000

On 13 Mar 2014 at 10:46:13, Gandalf Corvotempesta (gandalf.corvotempesta@xxxxxxxxx) wrote:

>
 Yes, if you have essentially high amount of commited data in the cluster 

>
 and/or large number of PG(tens of thousands). 

I've
 increased from 64 to 8192 PGs 

Do you mean you used PG splitting?
You should split PGs by a factor of 2x at a time. So to get from 64 to 8192, do 64->128,  then 128->256, …, 4096->8192.
Splitting is costly — it should be done with caution, even if you don’t have much data in the pools.

Cheers, Dan

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com