Hi Don,

On 03/03/2015 01:18, Don Doerner wrote:> Hello,
> 
>  
> 
> I am trying to set up to measure erasure coding performance and overhead.  My 
> Ceph “cluster-of-one” has 27 disks, hence 27 OSDs, all empty.  I have ots of 
> memory, and I am using “osd crush chooseleaf type = 0” in my config file, so 
> my OSDs should be able to peer with others on the same host, right?
> 
>  
> 
> I look at the EC profiles defined, and see only “default” which has k=2,m=1.  
> Wanting to set up a more realistic test, I defined a new profile “k8m3”, 
> similar to default, but with k=8,m=3. 
> 
>  
> 
> Checked with “ceph osd erasure-code-profile get k8m3”, all looks good.

When you create the erasure-code-profile you also need to set the failure 
domain (see ruleset-failure-domain in 
http://ceph.com/docs/master/rados/operations/erasure-code-jerasure/). It will 
not use the "osd crush chooseleaf type = 0" from your configuration file. You 
can verify the details of the ruleset used by the erasure coded pool with the 
command ./ceph osd crush rule dump

Cheers

> 
>  
> 
> I then go to define my pool: “ceph osd pool create ecpool 256 256 erasure 
> k8m3” apparently succeeds.
> 
> ·        Sidebar: my math on the pgnum stuff was (27 pools * 100)/11 = ~246, 
> round up to 256.
> 
>  
> 
> Now I ask “ceph health”, and get:
> 
> HEALTH_WARN256 pgs incomplete; 256 pgs stuck inactive; 256 pgs stuck unclean; 
> too few pgs per osd (9 < min 20)
> 
>  
> 
> Digging in to this a bit (“ceph health detail”), I see the magic OSD number 
> (2147483647) that says that there weren’t enough OSDs to assign to a 
> placement group, /for all placement groups/.  And at the same time, it is 
> warning me that I have too few PGs per OSD.
> 
>  
> 
> At the moment, I am defining a traditional replicated pool (3X) to see if 
> that will work…  Anyone have any guess as to what I may be doing incorrectly 
> with my erasure coded pool?  Or what I should do next to get a clue?
> 
>  
> 
> Regards,
> 
>  
> 
> -don-
> 
>  
> 
> ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
> The information contained in this transmission may be confidential. Any 
> disclosure, copying, or further distribution of confidential information is 
> not permitted unless such privilege is explicitly granted in writing by 
> Quantum. Quantum reserves the right to have electronic communications, 
> including email and attachments, sent across its networks filtered through 
> anti virus and spam software programs and retain such messages in order to 
> comply with applicable data security and retention requirements. Quantum is 
> not responsible for the proper and complete transmission of the substance of 
> this communication or for any delay in its receipt.
> 
> 
> _______________________________________________
> ceph-users mailing list
> [email protected]
> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
> 

-- 
Loïc Dachary, Artisan Logiciel Libre

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Reply via email to