Edit File: fslatency-nd.8
.TH fslatency-nd 8 "2015-01-30" "USER COMMANDS" .SH NAME fslatency-nd \- Measure file system sync read and write latency distribution. Uses Linux SystemTap (non-debuginfo). .SH SYNOPSIS .B fslatency-nd.stp [interval [count]] .SH DESCRIPTION fslatency-nd.stp dynamically traces two common file system functions: do_sync_read() and do_sync_write(), and reports a histogram distribution of latency. Many, but not all, file systems and workloads use these functions. Tracing their time provides one view of suffered file system latency. Other file system requests not covered by these functions and this tool include: asynchronous I/O, directory operations, and file handle operations. .SH REQUIREMENTS SystemTap. .SH OPTIONS .TP interval Interval for reports, in seconds. .TP count Number of reports to output. .SH EXAMPLES .TP Measure (most) file system synchronous read and write latency, and show a distribution histogram on Ctrl-C: # .B fslatency-nd.stp .TP Measure (most) file system synchronous read and write latency, and show a report every second, five times: # .B fslatency-nd.stp 1 5 .SH FIELDS .TP value The minimum value, in nanoseconds, for this histogram bucket. Read this value as "I/O latency was at least xxx". .TP count How many reads/writes fell into this latency range. .SH OVERHEAD The overheads are measurable, as file system requests can be very frequent, and the small overhead of tracing each (eg, 4 microseconds per call) can begin to add up. Best case, it won't be noticeable. Worst case, the application may run 2x slower. .SH SOURCE This is from the SystemTap lwtools collection. .IP https://github.com/brendangregg/systemtap-lwtools .PP Also look under the examples directory for a text file containing example usage, output, and commentary for this tool. .SH OS Linux .SH STABILITY Unstable - in development. .SH AUTHOR Brendan Gregg .SH SEE ALSO iostat(1), fsslower-nd.stp(8), biolatency-nd.stp(8)