Closed h-w-chen closed 2 years ago
When system is quite slow (approaching 4M records), one of the redis SI cpu utilization is high, as illustrated. The other 5 SI are having low cpu usage meanwhile (<2%)
this machine does not respond to ssh login; redis-cli is able to respond but quite slow (dbsize cmd delay is ~2.4s).
When system is quite slow (approaching 4M records), one of the redis SI cpu utilization is high, as illustrated. The other 5 SI are having low cpu usage meanwhile (<2%)
this machine does not respond to ssh login; redis-cli is able to respond but quite slow (dbsize cmd delay is ~2.4s).
possibly started memory thrashing
no issue before memory started to run out right?
nope; system is still running - though slowly
at the end of this run, please record the memory usage of rkv server too for reference
also disk usage of yaeger and ycsb
thanks for gathering those metrics.
this much used for index surprises me a bit.
closed; milestone has paased.
test config
redis persistence is disabled (save "")
records: 5M k-v payload: 5KB value
test procedure
load test only workloada setting
test result
ycsb log:
rkv memory usage when system is about 4M records
jaeger cpu and disk usage when system is about 4M recods
ycsb client cpu and disk usage when system is about 4M recods
observations