Hi,
> On 4 Oct 2022, at 03:36, Yuji Ito (伊藤 祐司) <yuji-ito@xxxxxxxxxxxx> wrote:
>
> After removing the index objects, I ran deep-scrub for all PGs of the index pool. However, the problem wasn't resolved.
Seems you just have large OMAPs, not because 'bogus shard' objects. Try to look PG stats with 'show_osd_pool_pg_usage.sh' script:
#!/bin/bash
OSD_POOL="$1"
if [ -z "${OSD_POOL}" ]
then
echo -e "Please provide osd pool name as first arg"
exit 1
fi
(header="id used_mbytes used_objects omap_used_mbytes omap_used_keys"
echo "${header}"
echo "${header}" | tr '[[:alpha:]_' '-'
ceph pg ls-by-pool "${OSD_POOL}" --format=json | jq -r '.pg_stats |
sort_by(.stat_sum.num_bytes) | .[] | (.pgid, .stat_sum.num_bytes/1024/1024,
.stat_sum.num_objects, .stat_sum.num_omap_bytes/1024/1024,
.stat_sum.num_omap_keys)' | paste - - - - -) | column -t
k
_______________________________________________
ceph-users mailing list -- ceph-users@xxxxxxx
To unsubscribe send an email to ceph-users-leave@xxxxxxx
