背景

centos7 在某一段時間監(jiān)控報警磁盤使用率達99%，由于監(jiān)控屬于概要形式信息，沒有快照信息的監(jiān)控(能發(fā)現(xiàn)某進程的I/O，CPU消耗情況)，所以需要在服務(wù)器上去定時執(zhí)行統(tǒng)計命令獲取快照信息。 需要通過iostat -dx -k去查看avgqu-sz、await、svctm、%util; sar -u查看%iowait、%user; pidstat -d 查看進程I/O讀寫的快照信息

步驟

• 生成統(tǒng)計信息文件

cat>/tmp/at_task.sh</tmp/pidstat_`date +%F_%T`.log 2>& 1 &
sar -u 2  >/tmp/sar_`date +%F_%T`.log 2>& 1 &
while [ 1 ];do echo -n `date +%T` >>/tmp/iostat_`date +%F` 2>& 1  && iostat -dx -k 1 1 >>/tmp/iostat_`date +%F` 2>& 1; sleep 2; done &
EOF

在while循環(huán)中使用iostat的原因是要輸出date +%T時間，不然只有數(shù)據(jù)，沒有時間信息也沒有什么用

• 使用at 命令定時執(zhí)行

at 15:14 today -f /tmp/at_task.sh

出現(xiàn)錯誤

Can't open /var/run/atd.pid to signal atd. No atd running?

重啟atd服務(wù)

service atd restart

重新開啟at定時任務(wù)

at 15:14 today -f /tmp/at_task.sh
job 2 at Wed Mar 13 1500 2019

得到如下快照信息
iostat

1535Linux 3.10.0-862.14.4.el7.x86_64 (ip-xxxxx)     03/13/2019      _x86_64_        (4 CPU)

Device:         rrqm/s   wrqm/s     r/s     w/s    rkB/s    wkB/s avgrq-sz avgqu-sz   await r_await w_await  svctm  %util
vda               0.12     0.07   17.31   19.41   580.79    90.52    36.57     0.09    2.39    4.42    0.57   0.72   2.63
scd0              0.00     0.00    0.00    0.00     0.00     0.00     6.00     0.00    0.28    0.28    0.00   0.25   0.00

sar

0300 PM     CPU     %user     %nice   %system   %iowait    %steal     %idle
0302 PM     all      0.25      0.00      0.38      0.00      0.00     99.37
0304 PM     all      1.25      0.13      0.63      0.00      0.00     97.99
0306 PM     all      0.25      0.13      0.50      0.00      0.00     99.12
0308 PM     all      0.50      0.00      0.50      0.63      0.00     98.37

pidstat

0300 PM   UID       PID   kB_rd/s   kB_wr/s kB_ccwr/s  Command
0302 PM  5700      9089      0.00      6.00      0.00  uxxx
0302 PM  5700      9140      0.00      6.00      0.00  uxxx
0302 PM  5700      9292      0.00     10.00      0.00  uxxx
0302 PM     0     18084      0.00      2.00      0.00  bash

kill 掉收集信息的命令

ps -ef | egrep 'iostat|sar|pidstat|while' | grep -v grep | awk '{print $2}' | xargs -l kill

但ps -ef | egrep 命令沒有獲取到while循環(huán)的pid，不kill掉該while循環(huán)，就會一直對/tmp/iostat_2019-03-13寫數(shù)據(jù)-_-

通過lsof 沒有定位到打開文件的進程

lsof /tmp/iostat_2019-03-13 
[root@ip-10-186-60-117 ~]#
[root@ip-10-186-60-117 ~]#

通過lsof 可以定位到打開mysql-error.log的進程

lsof /opt/mysql/data/5690/mysql-error.log 
COMMAND   PID                USER   FD   TYPE DEVICE SIZE/OFF     NODE NAME
mysqld  12858 actiontech-universe    1w   REG  253,1     6345 20083533 /opt/mysql/data/5690/mysql-error.log
mysqld  12858 actiontech-universe    2w   REG  253,1     6345 20083533 /opt/mysql/data/5690/mysql-error.log

可見，某進程只有一只持有某文件的inode，才可以通過lsof查看文件在被那些進程使用

獲取寫文件的進程號

安裝sysemtap

yum -y install systemtap

SystemTap 是對 Linux 內(nèi)核監(jiān)控和跟蹤的工具
利用systemtap中的inodewatch.stp工具來查找寫文件的進程號

得到文件的inode

stat -c '%i' /tmp/iostat_2019-03-13 
4210339

獲取文件所在設(shè)備的major，minor

ls -al /dev/vda1
brw-rw---- 1 root disk 253, 1 Jan 30 13:57 /dev/vda1

得到寫文件的pid

stap /usr/share/systemtap/examples/io/inodewatch.stp 253 1  4210339

Checking "/lib/modules/3.10.0-862.14.4.el7.x86_64/build/.config" failed with error: No such file or directory
Incorrect version or missing kernel-devel package, use: yum install kernel-devel-3.10.0-862.14.4.el7.x86_64

根據(jù)系統(tǒng)內(nèi)核版本在kernel-devel rpm build for :ScientificLinux7網(wǎng)站上下載相應(yīng)的kernal-devel包

wget ftp://ftp.pbone.net/mirror/ftp.scientificlinux.org/linux/scientific/7.2/x86_64/updates/security/kernel-devel-3.10.0-862.14.4.el7.x86_64.rpm

rpm -ivh kernel-devel-3.10.0-862.14.4.el7.x86_64.rpm

再次執(zhí)行stap

stap /usr/share/systemtap/examples/io/inodewatch.stp 253 1 4210339
......
Missing separate debuginfos, use: debuginfo-install kernel-3.10.0-862.14.4.el7.x86_64
Pass 2: analysis failed. [man error::pass2]
Number of similar error messages suppressed: 2.

安裝debuginfo kernal

debuginfo-install kernel-3.10.0-862.14.4.el7.x86_64
  Verifying  : kernel-debuginfo-common-x86_64-3.10.0-862.14.4.el7.x86_64                        1/3 
  Verifying  : yum-plugin-auto-update-debug-info-1.1.31-50.el7.noarch                           2/3 
  Verifying  : kernel-debuginfo-3.10.0-862.14.4.el7.x86_64                                      3/3 

Installed:
  kernel-debuginfo.x86_64 0:3.10.0-862.14.4.el7
  yum-plugin-auto-update-debug-info.noarch 0:1.1.31-50.el7

Dependency Installed:
  kernel-debuginfo-common-x86_64.x86_64 0:3.10.0-862.14.4.el7

再次執(zhí)行stap

stap /usr/share/systemtap/examples/io/inodewatch.stp 253 1  4210339
ERROR: module version mismatch (#1 SMP Tue Sep 25 1452 CDT 2018 vs #1 SMP Wed Sep 26 1511 UTC 2018), release 3.10.0-862.14.4.el7.x86_64
WARNING: /usr/bin/staprun exited with status: 1

添加 -v查看詳細報錯
stap -v  /usr/share/systemtap/examples/io/inodewatch.stp 253 1  4210339
Pass 1: parsed user script and 471 library scripts using 240276virt/41896res/3368shr/38600data kb, in 300usr/20sys/320real ms.
Pass 2: analyzed script: 2 probes, 12 functions, 8 embeds, 0 globals using 399436virt/196284res/4744shr/197760data kb, in 1540usr/560sys/2106real ms.
Pass 3: using cached /root/.systemtap/cache/f5/stap_f5c0cd780e8a2cac973c9e3ee69fba0c_7030.c
Pass 4: using cached /root/.systemtap/cache/f5/stap_f5c0cd780e8a2cac973c9e3ee69fba0c_7030.ko
Pass 5: starting run.
ERROR: module version mismatch (#1 SMP Tue Sep 25 1452 CDT 2018 vs #1 SMP Wed Sep 26 1511 UTC 2018), release 3.10.0-862.14.4.el7.x86_64
WARNING: /usr/bin/staprun exited with status: 1
Pass 5: run completed in 0usr/10sys/38real ms.
Pass 5: run failed.  [man error::pass5]

修改

vim /usr/src/kernels/3.10.0-862.14.4.el7.x86_64/include/generated/compile.h

#define UTS_VERSION "#1 SMP Tue Sep 25 1452 CDT 2018"
改為
#define UTS_VERSION "#1 SMP Wed Sep 26 1511 UTC 2018"

rm -rf  /root/.systemtap/cache/f5/stap_f5c0cd780e8a2cac973c9e3ee69fba0c_7030*

再次執(zhí)行

stap /usr/share/systemtap/examples/io/inodewatch.stp 253 1 4210339

iostat(4671) vfs_write 0xfd00001/4210339
iostat(4671) vfs_write 0xfd00001/4210339
iostat(4671) vfs_write 0xfd00001/4210339
iostat(4671) vfs_write 0xfd00001/4210339
iostat(4671) vfs_write 0xfd00001/4210339
iostat(4671) vfs_write 0xfd00001/4210339
iostat(4671) vfs_write 0xfd00001/4210339
iostat(4671) vfs_write 0xfd00001/4210339
iostat(4671) vfs_write 0xfd00001/4210339
iostat(4671) vfs_write 0xfd00001/4210339
iostat(4677) vfs_write 0xfd00001/4210339
iostat(4677) vfs_write 0xfd00001/4210339
iostat(4677) vfs_write 0xfd00001/4210339
iostat(4677) vfs_write 0xfd00001/4210339
iostat(4677) vfs_write 0xfd00001/4210339
iostat(4677) vfs_write 0xfd00001/4210339
iostat(4677) vfs_write 0xfd00001/4210339
iostat(4677) vfs_write 0xfd00001/4210339
iostat(4677) vfs_write 0xfd00001/4210339
iostat(4677) vfs_write 0xfd00001/4210339
iostat(4683) vfs_write 0xfd00001/4210339
............

可見已經(jīng)得到了寫/tmp/iostat_date +%F文件的進程號，但進程號一直在打印出來，因為后臺進程iostat -dx -m 的在while循環(huán)中的，每隔sleep 2s 后就會執(zhí)行一次iostat 產(chǎn)生新的pid。
那要怎樣才能讓iostat -dx -m 停止寫/tmp/iostat_date +%F文件了？除了重啟大法好 $_$

rm -rf 也不能終止后臺的while iostat進程寫文件，刪除了文件后，while循環(huán)又會生成新的文件

rm -rf  /tmp/iostat_2019-03-1*

stat /tmp/iostat_2019-03-1*
  File: ‘/tmp/iostat_2019-03-13’
  Size: 146700    Blocks: 512        IO Block: 4096   regular file
Device: fd01h/64769dInode: 4210339     Links: 1
Access: (0644/-rw-r--r--)  Uid: (    0/    root)   Gid: (    0/    root)
Access: 2019-03-14 1626.211888899 +0800
Modify: 2019-03-14 1617.854019793 +0800
Change: 2019-03-14 1617.854019793 +0800

正確做法

cat>/tmp/iostat.sh<while [ 1 ];do echo -n `date +%T` >>/tmp/iostat_`date +%F` 2>& 1  && iostat -dx -m 1 1 >>/tmp/iostat_`date +%F` 2>& 1; sleep 2; done &
EOF

at  now + 1 minute  today
bash /tmp/iostat.sh

#這樣就能方便的獲取到進程號pid了
ps -ef | grep iostat
root      8593     1  0 16:16 pts/2    0000 bash /tmp/iostat.sh