1btrfsslower(8) System Manager's Manual btrfsslower(8)
2
3
4
6 btrfsslower - Trace slow btrfs file operations, with per-event details.
7
9 btrfsslower [-h] [-j] [-p PID] [min_ms] [-d DURATION]
10
12 This tool traces common btrfs file operations: reads, writes, opens,
13 and syncs. It measures the time spent in these operations, and prints
14 details for each that exceeded a threshold.
15
16 WARNING: See the OVERHEAD section.
17
18 By default, a minimum millisecond threshold of 10 is used. If a thresh‐
19 old of 0 is used, all events are printed (warning: verbose).
20
21 Since this works by tracing the btrfs_file_operations interface func‐
22 tions, it will need updating to match any changes to these functions.
23
24 Since this uses BPF, only the root user can use this tool.
25
27 CONFIG_BPF and bcc.
28
30 -p PID Trace this PID only.
31
32 min_ms Minimum I/O latency (duration) to trace, in milliseconds.
33 Default is 10 ms.
34
35 -d DURATION
36 Total duration of trace in seconds.
37
39 Trace synchronous file reads and writes slower than 10 ms:
40 # btrfsslower
41
42 Trace slower than 1 ms:
43 # btrfsslower 1
44
45 Trace slower than 1 ms, and output just the fields in parsable format
46 (csv):
47 # btrfsslower -j 1
48
49 Trace all file reads and writes (warning: the output will be verbose):
50 # btrfsslower 0
51
52 Trace slower than 1 ms, for PID 181 only:
53 # btrfsslower -p 181 1
54
55 Trace for 10 seconds only:
56 # btrfsslower -d 10
57
59 TIME(s)
60 Time of I/O completion since the first I/O seen, in seconds.
61
62 COMM Process name.
63
64 PID Process ID.
65
66 T Type of operation. R == read, W == write, O == open, S == fsync.
67
68 OFF_KB File offset for the I/O, in Kbytes.
69
70 BYTES Size of I/O, in bytes.
71
72 LAT(ms)
73 Latency (duration) of I/O, measured from when it was issued by
74 VFS to the filesystem, to when it completed. This time is inclu‐
75 sive of block device I/O, file system CPU cycles, file system
76 locks, run queue latency, etc. It's a more accurate measure of
77 the latency suffered by applications performing file system I/O,
78 than to measure this down at the block device interface.
79
80 FILENAME
81 A cached kernel file name (comes from dentry->d_name.name).
82
83 ENDTIME_us
84 Completion timestamp, microseconds (-j only).
85
86 OFFSET_b
87 File offset, bytes (-j only).
88
89 LATENCY_us
90 Latency (duration) of the I/O, in microseconds (-j only).
91
93 This adds low-overhead instrumentation to btrfs writes and fsyncs, as
94 well as all system reads and opens (due to the current implementation
95 of the btrfs_file_operations interface). Particularly, all reads and
96 writes from the file system cache will incur extra overhead while trac‐
97 ing. Such reads and writes can be very frequent (depending on the work‐
98 load; eg, 1M/sec), at which point the overhead of this tool may become
99 noticeable. Measure and quantify before use. If this continues to be a
100 problem, consider switching to a tool that prints in-kernel summaries
101 only, such as btrfsdist(8).
102
103 Note that the overhead of this tool should be less than fileslower(8),
104 as this tool targets btrfs functions only, and not all file read/write
105 paths (which can include socket I/O).
106
108 This is from bcc.
109
110 https://github.com/iovisor/bcc
111
112 Also look in the bcc distribution for a companion _examples.txt file
113 containing example usage, output, and commentary for this tool.
114
116 Linux
117
119 Unstable - in development.
120
122 Brendan Gregg
123
125 btrfsdist(8), biosnoop(8), funccount(8), fileslower(8)
126
127
128
129USER COMMANDS 2016-02-15 btrfsslower(8)