From: Asias He <
as...@scylladb.com>
Committer: Avi Kivity <
a...@scylladb.com>
Branch: master
core: Reduce the default blocked-reactor-notify-ms to 25ms
A 200ms stall is way too much for most latency sensitive applications.
It makes more sense to use a smaller default number so the user will
know the stall and report.
Closes #1081
---
diff --git a/src/core/reactor.cc b/src/core/reactor.cc
--- a/src/core/reactor.cc
+++ b/src/core/reactor.cc
@@ -3474,7 +3474,7 @@ reactor_options::reactor_options(program_options::option_group* parent_group)
, task_quota_ms(*this, "task-quota-ms", 0.5, "Max time (ms) between polls")
, io_latency_goal_ms(*this, "io-latency-goal-ms", {}, "Max time (ms) io operations must take (1.5 * task-quota-ms if not set)")
, max_task_backlog(*this, "max-task-backlog", 1000, "Maximum number of task backlog to allow; above this we ignore I/O")
- , blocked_reactor_notify_ms(*this, "blocked-reactor-notify-ms", 200, "threshold in miliseconds over which the reactor is considered blocked if no progress is made")
+ , blocked_reactor_notify_ms(*this, "blocked-reactor-notify-ms", 25, "threshold in miliseconds over which the reactor is considered blocked if no progress is made")
, blocked_reactor_reports_per_minute(*this, "blocked-reactor-reports-per-minute", 5, "Maximum number of backtraces reported by stall detector per minute")
, blocked_reactor_report_format_oneline(*this, "blocked-reactor-report-format-oneline", true, "Print a simplified backtrace on a single line")
, relaxed_dma(*this, "relaxed-dma", "allow using buffered I/O if DMA is not available (reduces performance)")