mon: smooth io/recovery stats over longer period

On a busy cluster, the PGMap will turn over every 1-2 seconds and
OSDs will report PG stats every 5-10 seconds.  Averaging over only
two PGMap versions (2-4 seconds) will lead to erratic and spikey
delta values.  Move to 6 versions (6-12 seconds) to better match
the reporting interval while still keeping the values somewhat
responsive.

Signed-off-by: Sage Weil <sage@redhat.com>
This commit is contained in:
Sage Weil 2017-02-03 09:53:46 -05:00
parent 2d0dd51a9e
commit f285fa2ef2

View File

@ -282,7 +282,7 @@ OPTION(mon_osd_allow_primary_affinity, OPT_BOOL, false) // allow primary_affini
OPTION(mon_osd_prime_pg_temp, OPT_BOOL, true) // prime osdmap with pg mapping changes
OPTION(mon_osd_prime_pg_temp_max_time, OPT_FLOAT, .5) // max time to spend priming
OPTION(mon_osd_pool_ec_fast_read, OPT_BOOL, false) // whether turn on fast read on the pool or not
OPTION(mon_stat_smooth_intervals, OPT_INT, 2) // smooth stats over last N PGMap maps
OPTION(mon_stat_smooth_intervals, OPT_INT, 6) // smooth stats over last N PGMap maps
OPTION(mon_election_timeout, OPT_FLOAT, 5) // on election proposer, max waiting time for all ACKs
OPTION(mon_lease, OPT_FLOAT, 5) // lease interval
OPTION(mon_lease_renew_interval_factor, OPT_FLOAT, .6) // on leader, to renew the lease