11 Aug
2019
11 Aug
'19
11:06 p.m.
Hello everyone,
After discussing a few failure scenarios with different people on OFTC
network #ceph, I wanted to make sure that the following makes sense.
I have 2 Data Centers, please see attachment and I want to make sure
that if Data Center 1 (DC1) fails data recovery should be possible on
Data Center 2 (DC2)
As you can see on the diagram, I have 8 OSD's in total, 4 on each DC.
Using an Erasure Coded pool I've set the crush-failure-domain=rack.
- 2 DC's
- 8 OSD's in total
- EC pool profile used
crush-device-class=
crush-failure-domain=rack
crush-root=default
jerasure-per-chunk-alignment=false
k=2
m=2
plugin=jerasure
technique=reed_sol_van
w=8
Current crush layout:
# ceph osd df tree
ID CLASS WEIGHT REWEIGHT SIZE USE AVAIL %USE VAR PGS TYPE
NAME
-1 1.56238 - 1.6 TiB 280 GiB 1.3 TiB 17.47 1.00 - root
default
-22 0.39059 - 400 GiB 70 GiB 330 GiB 17.48 1.00 -
rack rack1-dc1
-3 0.19530 - 200 GiB 35 GiB 165 GiB 17.45 1.00
- host th-dev-ceph-osd-tc01
7 hdd 0.19530 1.00000 200 GiB 35 GiB 165 GiB 17.45 1.00
195 osd.7
-5 0.19530 - 200 GiB 35 GiB 165 GiB 17.50 1.00
- host th-dev-ceph-osd-tc02
1 hdd 0.19530 0.98708 200 GiB 35 GiB 165 GiB 17.50 1.00
195 osd.1
-20 0.39059 - 400 GiB 70 GiB 330 GiB 17.47 1.00 -
rack rack1-dc2
-13 0.19530 - 200 GiB 35 GiB 165 GiB 17.49 1.00
- host th-dev-ceph-osd-wp01
5 hdd 0.19530 1.00000 200 GiB 35 GiB 165 GiB 17.49 1.00
198 osd.5
-11 0.19530 - 200 GiB 35 GiB 165 GiB 17.45 1.00
- host th-dev-ceph-osd-wp02
0 hdd 0.19530 1.00000 200 GiB 35 GiB 165 GiB 17.45 1.00
197 osd.0
-19 0.39059 - 400 GiB 70 GiB 330 GiB 17.46 1.00 -
rack rack2-dc1
-7 0.19530 - 200 GiB 35 GiB 165 GiB 17.35 0.99
- host th-dev-ceph-osd-tc03
2 hdd 0.19530 1.00000 200 GiB 35 GiB 165 GiB 17.35 0.99
198 osd.2
-15 0.19530 - 200 GiB 35 GiB 165 GiB 17.57 1.01
- host th-dev-ceph-osd-tc04
6 hdd 0.19530 1.00000 200 GiB 35 GiB 165 GiB 17.57 1.01
198 osd.6
-21 0.39059 - 400 GiB 70 GiB 330 GiB 17.48 1.00 -
rack rack2-dc2
-9 0.19530 - 200 GiB 35 GiB 165 GiB 17.46 1.00
- host th-dev-ceph-osd-wp03
3 hdd 0.19530 1.00000 200 GiB 35 GiB 165 GiB 17.46 1.00
197 osd.3
-17 0.19530 - 200 GiB 35 GiB 165 GiB 17.50 1.00
- host th-dev-ceph-osd-wp04
4 hdd 0.19530 1.00000 200 GiB 35 GiB 165 GiB 17.50 1.00
198 osd.4
TOTAL 1.6 TiB 280 GiB 1.3 TiB
17.47
MIN/MAX VAR: 0.99/1.01 STDDEV: 0.06
I am aware that this is not optimal so please advise. What EC profile
should I chose, what should k and m values be and should I use LRC
instead of JERASURE algorithm considering the attached jpeg.
Thank you very much in advance!
--
Met vriendelijke groeten, Kind regards,
Valentin Bajrami
Target Holding
