ceph gps backfill_toofull

目的

修复 backfill_toolfull 故障

故障显示如下

 ceph -c  -s
  cluster:
    id:     axxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
    health: HEALTH_WARN
            6 clients failing to respond to cache pressure
            Low space hindering backfill (add storage if this doesn't resolve itself): 2 pgs backfill_toofull

具体信息

# ceph health detail
[WRN] PG_BACKFILL_FULL: Low space hindering backfill (add storage if this doesn't resolve itself): 2 pgs backfill_toofull
    pg 12.5c9 is active+remapped+backfill_wait+backfill_toofull, acting [125,165,196]
    pg 12.dd1 is active+remapped+backfill_wait+backfill_toofull, acting [79,59,104]

查询对应 OSD 空间

$ ceph osd df | grep -E '^123|^\ 79|^165|^196|^\ 59|^104'
ID   CLASS  WEIGHT    REWEIGHT  SIZE     RAW USE  DATA     OMAP     META     AVAIL    %USE   VAR   PGS  STATUS
104    ssd  13.97069   1.00000   14 TiB  5.6 TiB  5.6 TiB  329 MiB   11 GiB  8.3 TiB  40.27  0.95  105      up
196    ssd  14.00000   1.00000   14 TiB  5.6 TiB  5.6 TiB  388 MiB   12 GiB  8.4 TiB  39.87  0.94   90      up
 59    ssd  13.97069   1.00000   14 TiB  4.8 TiB  4.8 TiB  209 MiB   10 GiB  9.1 TiB  34.54  0.81   86      up
165    ssd  14.00000   1.00000   14 TiB  5.8 TiB  5.8 TiB  298 MiB   10 GiB  8.2 TiB  41.37  0.97  112      up
 79    ssd  13.97069   1.00000   14 TiB  5.6 TiB  5.6 TiB  271 MiB   11 GiB  8.3 TiB  40.31  0.95  114      up
125    ssd  14.00000   1.00000   14 TiB  5.6 TiB  5.6 TiB  406 MiB   11 GiB  8.3 TiB  40.43  0.95  123      up

目测已用空间不到50%， 先观察