Home

jc.parsers.nfsiostat_s

jc - JSON Convert nfsiostat command output streaming parser

This streaming parser outputs JSON Lines (cli) or returns an Iterable of Dictionaries (module)

Parses the per-mount reports printed by nfsiostat from nfs-utils. Use it with an interval and no count (e.g. nfsiostat 1), which runs until it is stopped.

Each report is converted to one object per mount point. A mount object is output when the next mount header is read or the input ends, so with an interval the last mount of each report is output when the next report starts.

With an interval, the first report for each mount covers the time since the file system was mounted, and later reports cover each interval.

The cache summary lines printed by the -a, -d, and -p options are not parsed.

Usage (cli):

$ nfsiostat 1 | jc --nfsiostat-s

Note: When piping jc converted nfsiostat output to other processes it may appear the output is hanging due to the OS pipe buffers. This is because nfsiostat output is too small to quickly fill up the buffer. Use the -u option to unbuffer the jc output if you would like immediate output. See the readme for more information.

Usage (module):

import jc

result = jc.parse('nfsiostat_s', nfsiostat_command_output.splitlines())
for item in result:
    # do something

Schema:

{
  "filesystem":               string,
  "mount":                    string,
  "ops_sec":                  float,
  "rpc_bklog":                float,
  "read": {
    "ops_sec":                float,
    "kb_sec":                 float,
    "kb_op":                  float,
    "retrans":                integer,
    "retrans_percent":        float,
    "avg_rtt_ms":             float,
    "avg_exe_ms":             float,
    "avg_queue_ms":           float,
    "errors":                 integer,    # [2]
    "errors_percent":         float       # [2]
  },
  "write": {
    "ops_sec":                float,
    "kb_sec":                 float,
    "kb_op":                  float,
    "retrans":                integer,
    "retrans_percent":        float,
    "avg_rtt_ms":             float,
    "avg_exe_ms":             float,
    "avg_queue_ms":           float,
    "errors":                 integer,    # [2]
    "errors_percent":         float       # [2]
  },
  "getattr": {                # [0]
    "ops_sec":                float,
    "kb_sec":                 float,
    "kb_op":                  float,
    "retrans":                integer,
    "retrans_percent":        float,
    "avg_rtt_ms":             float,
    "avg_exe_ms":             float,
    "avg_queue_ms":           float,
    "errors":                 integer,    # [2]
    "errors_percent":         float       # [2]
  },
  "access": {                 # [0]
    "ops_sec":                float,
    "kb_sec":                 float,
    "kb_op":                  float,
    "retrans":                integer,
    "retrans_percent":        float,
    "avg_rtt_ms":             float,
    "avg_exe_ms":             float,
    "avg_queue_ms":           float,
    "errors":                 integer,    # [2]
    "errors_percent":         float       # [2]
  },
  "lookup": {                 # [1]
    "ops_sec":                float,
    "kb_sec":                 float,
    "kb_op":                  float,
    "retrans":                integer,
    "retrans_percent":        float,
    "avg_rtt_ms":             float,
    "avg_exe_ms":             float,
    "avg_queue_ms":           float,
    "errors":                 integer,    # [2]
    "errors_percent":         float       # [2]
  },
  "readdir": {                # [1]
    "ops_sec":                float,
    "kb_sec":                 float,
    "kb_op":                  float,
    "retrans":                integer,
    "retrans_percent":        float,
    "avg_rtt_ms":             float,
    "avg_exe_ms":             float,
    "avg_queue_ms":           float,
    "errors":                 integer,    # [2]
    "errors_percent":         float       # [2]
  },
  "readdirplus": {            # [1]
    "ops_sec":                float,
    "kb_sec":                 float,
    "kb_op":                  float,
    "retrans":                integer,
    "retrans_percent":        float,
    "avg_rtt_ms":             float,
    "avg_exe_ms":             float,
    "avg_queue_ms":           float,
    "errors":                 integer,    # [2]
    "errors_percent":         float       # [2]
  },

  # below object only exists if using -qq or ignore_exceptions=True
  "_jc_meta": {
    "success":                boolean,    # false if error parsing
    "error":                  string,     # exists if "success" is false
    "line":                   string      # exists if "success" is false
  }
}

[0] only with the `-a` option
[1] only with the `-d` option. `readdirplus` only appears for NFS
    versions that have a READDIRPLUS operation (NFSv3)
[2] only present when the kernel reports per-operation error counts

Examples:

$ nfsiostat 1 | jc --nfsiostat-s
{"filesystem":"192.168.1.10:/export/home","mount":"/home","ops_sec":2.11,...}
{"filesystem":"nas01.example.com:/volume1/archive","mount":"/mnt/archive",...}
{"filesystem":"192.168.1.10:/export/home","mount":"/home","ops_sec":1450.0,...}
...

$ nfsiostat 1 | jc --nfsiostat-s -r
{"filesystem":"192.168.1.10:/export/home","mount":"/home","ops_sec":"2.110",...}
{"filesystem":"nas01.example.com:/volume1/archive","mount":"/mnt/archive",...}
{"filesystem":"192.168.1.10:/export/home","mount":"/home","ops_sec":"1450.000",...}
...

parse

def parse(
    data: Iterable[str],
    raw: bool = False,
    quiet: bool = False,
    ignore_exceptions: bool = False
) -> Iterator[Dict[str, Any] | Tuple[BaseException, str]]

Main text parsing generator function. Returns an iterable object.

Parameters:

data:              (iterable)  line-based text data to parse
                               (e.g. sys.stdin or str.splitlines())

raw:               (boolean)   unprocessed output if True
quiet:             (boolean)   suppress warning messages if True
ignore_exceptions: (boolean)   ignore parsing exceptions if True

Returns:

Iterable of Dictionaries

Parser Information

Compatibility: linux

Source: jc/parsers/nfsiostat_s.py

Version 1.0 by Yang Bai (271969568+ybai08@users.noreply.github.com)