Ceph » CephFS

Through testing the bash script for MDS cluster, create a test directory under the ceph mount path, in the test directory to create 100000 small file, when the script quickly after writing 10000 files, ceph cluster alarm: "No space left on device", in the Google directory can be solved by setting the parameters in the query problem of division
Operation:
cat /etc/ceph/ceph.conf
osd crush location hook = /usr/bin/calamari-crush-location
public_network = 10.0.0.0/24
osd_pool_default_size = 3
osd_pool_default_min_size = 2
osd_pool_default_pgp_num = 256
osd_pool_default_pg_num = 256
osd_crush_chooseleaf_type = 1
mon_clock_drift_allowed = 2
mon_clock_drift_warn_backoff = 30
hellomon_osd_full_ratio = .85
mon_osd_nearfull_ratio = .85
debug rgw = 20
#rgw_override_bucket_index_max_shards = 0
mds_bal_frag = true
#mon_pg_warn_max_per_osd = 0

command：

Ceph fs set dbleader allow_dirfrags true - yes - - really - I mean - it

+++++++++++++++++++++++++++++++++ script +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

#!/bin/bash
let START=`date +'%s'`
for i in {1..400000};do
(( mod=$i % 1000 ))
if (( mod == 0 )); then
END=`date +'%s'`
let elapse=END-START
echo $i, $elapse
fi
dd if=/dev/zero of=./new-$i-new1 bs=200k count=1
done
END=`date +'%s'`
let elapse=END-START
echo $START, $END, $elapse