6 Dec
2019
6 Dec
'19
2:01 p.m.
On 6.12.19 14:57, Jochen Schulz wrote:
Hi!
Thank you!
The output of both commands are below.
I still dont understand why there are 21T used data (because 5.5T*3 =
16.5T != 21T) and why there seems to be only 4.5 T MAX AVAIL, but the
osd output tells we have 25T free space.
As I know MAX AVAIL is calculated with respect to
mon_osd_full_ratio, max OSD %USE, WEIGHT, total weight of set of OSDs in
pool and replication factor of pool.
Max used hdd OSD is 34.
So (0.95 - 0.65) * 931 / 3 * 1.3 ~= 4.5T
1.3 here state for avg osd size / 931 (you have 13*1.8T osd and other
24*931G)
The logic behind is to estimate amount of data you can write into pool
before max used OSD became full.
Sorry, can't say anything precise about 82% and backfill_toofull pg
while max used OSD is 65% used.