summaryrefslogtreecommitdiff
path: root/doc/administration/troubleshooting/gdb-stuck-ruby.txt
blob: 13d5dfcffa4ca3002826f87ddec432ac2422edcd (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
# Here's the script I'll use to demonstrate - it just loops forever:

$ cat test.rb 
#!/usr/bin/env ruby

loop do
  sleep 1
end

# Now, I'll start the script in the background, and redirect stdout and stderr
# to /dev/null:

$ ruby ./test.rb >/dev/null 2>/dev/null &
[1] 1343

# Next, I'll grab the PID of the script (1343):

$ ps aux | grep test.rb
vagrant   1343  0.0  0.4   3884  1652 pts/0    S    14:42   0:00 ruby ./test.rb
vagrant   1345  0.0  0.2   4624   852 pts/0    S+   14:42   0:00 grep --color=auto test.rb

# Now I start gdb. Note that I'm using sudo here. This may or may not be
# necessary in your setup. I'd try without sudo first, and fall back to adding
# it if the next step fails:

$ sudo gdb
GNU gdb (Ubuntu/Linaro 7.4-2012.04-0ubuntu2.1) 7.4-2012.04
Copyright (C) 2012 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
and "show warranty" for details.
This GDB was configured as "i686-linux-gnu".
For bug reporting instructions, please see:
<http://bugs.launchpad.net/gdb-linaro/>.

# OK, now I'm in gdb, and I want to instruct it to attach to our Ruby process.
# I can do that using the 'attach' command, which takes a PID (the one we
# gathered above):

(gdb) attach 1343
Attaching to process 1343
Reading symbols from /opt/vagrant_ruby/bin/ruby...done.
Reading symbols from /lib/i386-linux-gnu/librt.so.1...(no debugging symbols found)...done.
Loaded symbols for /lib/i386-linux-gnu/librt.so.1
Reading symbols from /lib/i386-linux-gnu/libdl.so.2...(no debugging symbols found)...done.
Loaded symbols for /lib/i386-linux-gnu/libdl.so.2
Reading symbols from /lib/i386-linux-gnu/libcrypt.so.1...(no debugging symbols found)...done.
Loaded symbols for /lib/i386-linux-gnu/libcrypt.so.1
Reading symbols from /lib/i386-linux-gnu/libm.so.6...(no debugging symbols found)...done.
Loaded symbols for /lib/i386-linux-gnu/libm.so.6
Reading symbols from /lib/i386-linux-gnu/libc.so.6...(no debugging symbols found)...done.
Loaded symbols for /lib/i386-linux-gnu/libc.so.6
Reading symbols from /lib/i386-linux-gnu/libpthread.so.0...(no debugging symbols found)...done.
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/i386-linux-gnu/libthread_db.so.1".
Loaded symbols for /lib/i386-linux-gnu/libpthread.so.0
Reading symbols from /lib/ld-linux.so.2...(no debugging symbols found)...done.
Loaded symbols for /lib/ld-linux.so.2
0xb770c424 in __kernel_vsyscall ()

# Great, now gdb is attached to the target process. If the step above fails, try
# going back and running gdb under sudo. The next thing I want to do is gather
# C-level backtraces from all threads in the process. The following command
# stands for 'thread apply all backtrace':

(gdb) t a a bt

Thread 1 (Thread 0xb74d76c0 (LWP 1343)):
#0  0xb770c424 in __kernel_vsyscall ()
#1  0xb75d7abd in select () from /lib/i386-linux-gnu/libc.so.6
#2  0x08069c56 in rb_thread_wait_for (time=...) at eval.c:11376
#3  0x080a20fd in rb_f_sleep (argc=1, argv=0xbf85f490) at process.c:1633
#4  0x0805e0e2 in call_cfunc (argv=0xbf85f490, argc=1, len=-1, recv=3075299660, func=0x80a20b0 <rb_f_sleep>)
    at eval.c:5778
#5  rb_call0 (klass=3075304600, recv=3075299660, id=9393, oid=9393, argc=1, argv=0xbf85f490, body=0xb74c85a8, flags=2)
    at eval.c:5928
#6  0x0805e35d in rb_call (klass=3075304600, recv=3075299660, mid=9393, argc=1, argv=0xbf85f490, scope=1, 
    self=<optimized out>) at eval.c:6176
#7  0x080651ec in rb_eval (self=3075299660, n=0xb74c4e1c) at eval.c:3521
#8  0x0805c31c in rb_yield_0 (val=6, self=3075299660, klass=<optimized out>, flags=0, avalue=0) at eval.c:5095
#9  0x0806a1e5 in loop_i () at eval.c:5227
#10 0x08058dbd in rb_rescue2 (b_proc=0x806a1c0 <loop_i>, data1=0, r_proc=0, data2=0) at eval.c:5491
#11 0x08058f28 in rb_f_loop () at eval.c:5252
#12 0x0805e0c1 in call_cfunc (argv=0x0, argc=0, len=0, recv=3075299660, func=0x8058ef0 <rb_f_loop>) at eval.c:5781
#13 rb_call0 (klass=3075304600, recv=3075299660, id=4121, oid=4121, argc=0, argv=0x0, body=0xb74d4dbc, flags=2)
    at eval.c:5928
#14 0x0805e35d in rb_call (klass=3075304600, recv=3075299660, mid=4121, argc=0, argv=0x0, scope=1, self=<optimized out>)
    at eval.c:6176
#15 0x080651ec in rb_eval (self=3075299660, n=0xb74c4dcc) at eval.c:3521
#16 0x080662c6 in rb_eval (self=3075299660, n=0xb74c4de0) at eval.c:3236
#17 0x08068ee4 in ruby_exec_internal () at eval.c:1654
#18 0x08068f24 in ruby_exec () at eval.c:1674
#19 0x0806b2cd in ruby_run () at eval.c:1684
#20 0x08053771 in main (argc=2, argv=0xbf860204, envp=0xbf860210) at main.c:48

# C backtraces are sometimes sufficient, but often Ruby backtraces are necessary
# for debugging as well. Ruby has a built-in function called rb_backtrace() that
# we can use to dump out a Ruby backtrace, but it prints to stdout or stderr
# (depending on your Ruby version), which might have been redirected to a file
# or to /dev/null (as in our example) when the process started up.
#
# To get aroundt this, we'll do a little trick and redirect the target process's
# stdout and stderr to the current TTY, so that any output from the process
# will appear directly on our screen.

# First, let's close the existing file descriptors for stdout and stderr
# (FD 1 and 2, respectively):
(gdb) call (void) close(1)
(gdb) call (void) close(2)

# Next, we need to figure out the device name for the current TTY:
(gdb) shell tty
/dev/pts/0

# OK, now we can pass the device name obtained above to open() and attach
# file descriptors 1 and 2 back to the current TTY with these calls:

(gdb) call (int) open("/dev/pts/0", 2, 0)
$1 = 1
(gdb) call (int) open("/dev/pts/0", 2, 0)
$2 = 2

# Finally, we call rb_backtrace() in order to dump the Ruby backtrace:

(gdb) call (void) rb_backtrace()
  from ./test.rb:4:in `sleep'
  from ./test.rb:4
  from ./test.rb:3:in `loop'
  from ./test.rb:3

# And here's how we get out of gdb. Once you've quit, you'll probably want to
# clean up the stuck process by killing it.

(gdb) quit
A debugging session is active.

  Inferior 1 [process 1343] will be detached.

Quit anyway? (y or n) y
Detaching from program: /opt/vagrant_ruby/bin/ruby, process 1343
$