SELECT * in MemSQL -
i playing memsql , seems fast. when execute select * [some big table]
takes long time. can see lot of traffic in memsql opt , no cpu/memory usage. have set sql editor (datagrip) fetch 500 rows doesn't (i know can use limit command).
my question going on? of partitions going stream of results client fetch 500 rows after that? there way how monitor this?
what's going on situation dependent, can use explain select * [some big table]
, should tell query doing in broad sense can give indication step in query execution taking long time
if it's not going stream of results client client adding limit you. can use show plancache
, presence of select * [some big table] limit 500
see if client adding automatically. if running query limit should going faster, not surprise me if truncates results on end - won't improve speed @ all.
if monitor cpu, memory, disk i/o, , network you've got covered. unless you're using column store shouldn't using disk @ all, there's either error or network bottleneck. culprit rough distributed join, can quite slow - in older versions.
Comments
Post a Comment