NAME

       cachetop  -  Statistics  for  linux  page  cache hit/miss ratios per processes. Uses Linux
       eBPF/bcc.

SYNOPSIS

       cachetop [-p PID] [interval]

DESCRIPTION

       This traces four kernel  functions  and  prints  per-processes  summaries  every  interval
       seconds.  This  can  be  useful  for  processes workload characterization, and looking for
       patterns in operation usage over time. It provides a top-like interface which  by  default
       sorts by HITS in ascending order.

       This  works  by  tracing  kernel page cache functions using dynamic tracing, and will need
       updating to match any changes to these functions.  Edit  the  script  to  customize  which
       functions are traced.

       Since this uses BPF, only the root user can use this tool.

OPTIONS

       -p PID Trace this PID only.

KEYBINDINGS

       The following keybindings can be used to control the output of cachetop.

       <      Use the previous column for sorting.

       >      Use the next column for sorting.

       r      Toggle sorting order (default ascending).

       q      Quit cachetop.

REQUIREMENTS

       CONFIG_BPF and bcc.

EXAMPLES

       Update summaries every five second:
              # cachetop

       Print summaries each second:
              # cachetop 1

FIELDS

       PID    Process ID of the process causing the cache activity.

       UID    User ID of the process causing the cache activity.

       CMD    Name of the process.

       HITS   Number of page cache hits.

       MISSES Number of page cache misses.

       DIRTIES
              Number of dirty pages added to the page cache.

       READ_HIT%
              Read hit percent of page cache usage.

       WRITE_HIT%
              Write hit percent of page cache usage.

       BUFFERS_MB
              Buffers size taken from /proc/meminfo.

       CACHED_MB
              Cached amount of data in current page cache taken from /proc/meminfo.

OVERHEAD

       This  traces various kernel page cache functions and maintains in-kernel counts, which are
       asynchronously copied to user-space. While  the  rate  of  operations  can  be  very  high
       (>1G/sec)  we  can  have  up  to 34% overhead, this is still a relatively efficient way to
       trace these events, and so the overhead is expected to  be  small  for  normal  workloads.
       Measure in a test environment.

SOURCE

       This is from bcc.

              https://github.com/iovisor/bcc

       Also  look  in  the bcc distribution for a companion _examples.txt file containing example
       usage, output, and commentary for this tool.

OS

       Linux

STABILITY

       Unstable - in development.

AUTHOR

       Emmanuel Bretelle, Rocky Xing

SEE ALSO

       cachestat (8)