~/talk

Elastic Scheduling with TiKV

name:Song Gao
links:none on file
links:none on file
slides:none on file

Elastic scheduling adds an operator + PD-based pipeline that auto-scales TiKV on Kubernetes: TiKV exports metrics to Prometheus, PD pulls metrics and runs a hot-region scheduler that migrates only high read/write regions via Raft leader transfer and replica add/remove, and an operator CRD/controller labels new nodes (e.g. special_use=hot-region) so only hot schedulers target them. A sysbench/TiDB experiment spins up two nodes, moves hot regions and shows TPS/QPS increase; future work adds follower-read replicas and cold-to-cheap-storage tiering.

AI-written summary · about

  1. 0:00Start
  2. 0:59Agenda
  3. 1:47Background Information
  4. 2:28Architecture
  5. 6:09Elastic Scheduling
  6. 7:10Why Do We Need or Choose the Elastic Scheduling
  7. 11:37Operator Pattern
  8. 16:45Label Filter
  9. 17:12Elastic Scheduling Configuration
~/search
type to search 6,249 conferences, 138,987 speakers and 185,172 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.