诊断Java线程问题时，一直要求提供客户Java的线程堆。大部分时间都在指导客户怎么抓数据，非常浪费时间。曾经写了一个非常方便的脚本，但是用户一般都会纠结于参数的选择，效果不明显。
有兴趣的可以使用我的脚本 ajs2.sh
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
#!/bin/bash
#--------------------------------------------------------------------------------
#- File name: ajs.sh - (A)uto (J)ava (S)tack
#- Purpose: Collect Java Stacks automatically
#- Author: jie.chen@oracle.com
#- Usage: ./ajs.sh -p <process keyword> [interval in second] [times]
#--------------------------------------------------------------------------------
version=0.1
showApp(){
echo -e "\t\t(A)uto (J)ava (S)tack\t(v"$version")"
}
showHelp(){
echo
showApp
echo "Syntax:"
echo -e "\t"$(basename $0)" -p <process keyword> [interval in second] [times]"
echo "Example:"
echo -e "\t1) for weblogic cluster managed: "$(basename $0)" -p myweblogic-ManagedServer1 5 8"
echo -e "\t2) for weblogic standlone: "$(basename $0)" -p weblogic.Server"
echo -e "\t3) for apache tomcat: "$(basename $0)" -p catalina 5 8"
echo -e "\t4) for common java: "$(basename $0)" -p 1234 5 8"
echo
#exit
}
# default parameters
pname="java"
op="-p"
args=(5 8)
trcfile=$(basename $0)
pfilter=grep # to filter out "grep" itself
## read parameters from shell command line
while getopts HhP:p: opt
do
case "$opt" in
P | p) pname=$OPTARG
op="-p";;
H | h) showHelp
exit;;
*) ;;
esac
done
i=0
shift $[ $OPTIND - 1 ]
for param in "$@"
do
args[$i]=$param
i=$i+1
done
tsleep=${args[0]}
ttimes=${args[1]}
getPDesc(){
if [ -z "$pfilter" ]
then
ps -ef | grep $pname | grep -v $(basename $0)
else
ps -ef | grep $pname | grep -v $(basename $0) | grep -v $pfilter
fi
}
pdesc=`getPDesc`
if [ -z "$pdesc" ]
then
echo "!!! No matched process '$pname' found !!!"
echo "Type command '$(basename $0) -h' for help"
#showHelp
exit
fi
pid=`echo $pdesc | cut -d ' ' -f 2`
trcfile=$trcfile"_"$pid".txt"
cat > $trcfile << EOF # new file
`showApp`
--------------------------------------------------------------
$(basename $0) $op $pname $tsleep $ttimes
$pdesc
EOF
echo "PID" $pid "found" | tee -a $trcfile ## append file
for ((i=1; i<=$ttimes; i++))
do
printf "[%-40s] %d%%\r" 'Collecting Java Stacks ... ' $[100/$ttimes*$i]
echo '-------------- Thread Dumps '$i >> $trcfile
jstack $pid >> $trcfile
#jmap -heap $pid >> $trcfile
if [ $i -ne $ttimes ]
then
sleep $tsleep
fi
done
echo '-------------- JVM Env Settings' >> $trcfile
jinfo $pid >> $trcfile
printf "[%-40s] %d%%" 'Collected ... ' $[100]
echo
echo "Dumped Java Stack Traces File: "`pwd`"/"$trcfile
少就是多=简洁
上面我已经尽量地让程序运行傻瓜化，但也一直在思考把工具写得更加简单，最近构造出一行简短的代码，让一些用户试了一下，效果很好。
1
pid=30994;t=5;jinfo $pid >$pid & watch -n $t "grep -c 'Full thread dump Java HotSpot(TM)' $pid;jstack $pid >> $pid"
其中：
- pid：Java Process ID
- t：抓取的秒数间隔
执行结果：
1
2
3
Every 5.0s: grep -c 'Full thread dump Java HotSpot(TM)' 30994;jstack 30994 >> 30994 Sat Apr 27 22:52:39 2019
8
最后一个数字 8 代表当前时间已经抓取到的线程堆次数。可以根据实际问题决定何时退出程序。