Erasure Code Setup

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Folks-

 

Having a bit of trouble with EC setup on 0.78.  Hoping someone can help me out.  I’ve got most of the pieces in place, I think I’m just having a problem with the ruleset.

 

I am running 0.78:

ceph --version

ceph version 0.78 (f6c746c314d7b87b8419b6e584c94bfe4511dbd4)

 

I created a new ruleset:

ceph osd crush rule create-erasure ecruleset

 

Then I created a new erasure code pool:

ceph osd pool create mycontainers_1 1800 1800 erasure crush_ruleset=ecruleset erasure-code-k=9 erasure-code-m=3

 

Pool exists:

ceph@joceph-admin01:/etc/ceph$ ceph osd dump

epoch 106

fsid b12ebb71-e4a6-41fa-8246-71cbfa09fb6e

created 2014-03-24 12:06:28.290970

modified 2014-03-24 12:42:59.231381

flags

pool 0 'data' replicated size 1 min_size 1 crush_ruleset 0 object_hash rjenkins pg_num 128 pgp_num 128 last_change 84 owner 0 flags hashpspool crash_replay_interval 45 stripe_width 0

pool 1 'metadata' replicated size 1 min_size 1 crush_ruleset 0 object_hash rjenkins pg_num 128 pgp_num 128 last_change 86 owner 0 flags hashpspool stripe_width 0

pool 2 'rbd' replicated size 1 min_size 1 crush_ruleset 0 object_hash rjenkins pg_num 128 pgp_num 128 last_change 88 owner 0 flags hashpspool stripe_width 0

pool 4 'mycontainers_2' replicated size 2 min_size 2 crush_ruleset 0 object_hash rjenkins pg_num 1200 pgp_num 1200 last_change 100 owner 0 flags hashpspool stripe_width 0

pool 5 'mycontainers_3' replicated size 1 min_size 1 crush_ruleset 0 object_hash rjenkins pg_num 1800 pgp_num 1800 last_change 94 owner 0 flags hashpspool stripe_width 0

pool 6 'mycontainers_1' erasure size 12 min_size 1 crush_ruleset 1 object_hash rjenkins pg_num 1800 pgp_num 1800 last_change 104 owner 0 flags hashpspool stripe_width 4320

 

However, the new PGs won’t come to a healthy state:

ceph@joceph-admin01:/etc/ceph$ ceph status

    cluster b12ebb71-e4a6-41fa-8246-71cbfa09fb6e

     health HEALTH_WARN 1800 pgs incomplete; 1800 pgs stuck inactive; 1800 pgs stuck unclean

     monmap e1: 2 mons at {mohonpeak01=10.0.0.101:6789/0,mohonpeak02=10.0.0.102:6789/0}, election epoch 4, quorum 0,1 mohonpeak01,mohonpeak02

     osdmap e106: 18 osds: 18 up, 18 in

      pgmap v261: 5184 pgs, 7 pools, 0 bytes data, 0 objects

            682 MB used, 15082 GB / 15083 GB avail

                3384 active+clean

                1800 incomplete

 

I think this is because it is using a failure domain of hosts and I only have 2 hosts (with 9 OSDs on each for 18 OSDs total).  I suspect I need to change the ruleset to use a failure domain of OSD instead of host.  This is also mentioned on this page: https://ceph.com/docs/master/dev/erasure-coded-pool/.

 

However, the guidance on that that page to adjust it using commands of the form “ceph osd erasure-code-profile set myprofile” is not working for me.  As far as I can tell “ceph osd erasure-code-profile” does not seem to be a valid command syntax.  Is this documentation correct and up to date for 0.78?  Can anyone suggest where I am going wrong?  Thanks!

 

ceph@joceph-admin01:/etc/ceph$ ceph osd erasure-code-profile ls

no valid command found; 10 closest matches:

osd tier add-cache <poolname> <poolname> <int[0-]>

osd tier set-overlay <poolname> <poolname>

osd tier remove-overlay <poolname>

osd tier remove <poolname> <poolname>

osd tier cache-mode <poolname> none|writeback|forward|readonly

osd thrash <int[0-]>

osd tier add <poolname> <poolname> {--force-nonempty}

osd stat

osd reweight-by-utilization {<int[100-]>}

osd pool stats {<name>}

Error EINVAL: invalid command

ceph@joceph-admin01:/etc/ceph$

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

[Index of Archives]     [Information on CEPH]     [Linux Filesystem Development]     [Ceph Development]     [Ceph Large]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [xfs]


  Powered by Linux