文件系统分析工具我们也分为传统工具,BPF和System三个分别进行介绍,我来为大家科普一下关于操作系统文件管理架构图?以下内容希望对你有帮助!
操作系统文件管理架构图
文件系统分析工具我们也分为传统工具,BPF和System三个分别进行介绍。
传统工具dfdf通常可以查看文件系统的使用情况,如:
~# df -h
Filesystem Size Used Avail Use% Mounted on
devtmpfs 1.9G 0 1.9G 0% /dev
tmpfs 1.9G 0 1.9G 0% /dev/shm
tmpfs 1.9G 41M 1.9G 3% /run
tmpfs 1.9G 0 1.9G 0% /sys/fs/cgroup
/dev/sda1 50G 6.6G 44G 14% /
/dev/sdb 20G 45M 19G 1% /data
tmpfs 379M 0 379M 0% /run/User/1000
tmpfs 379M 0 379M 0% /run/user/0
可以关注Use%这一列,因为在文件系统使用率较高时会导致文件系统的性能下降,由于资源碎片等原因会将顺序I/O降级为随机I/O。但这也不是绝对的,需要看具体的文件系统的实现方式。
mountmout命令可以将文件系统挂载到系统上,也可以列出这些文件系统的类型和挂载参数:
~# mount
sysfs on /sys type sysfs (rw,nosuid,nodev,noexec,relatime,seclabel)
proc on /proc type proc (rw,nosuid,nodev,noexec,relatime)
devtmpfs on /dev type devtmpfs (rw,nosuid,seclabel,size=1915812k,nr_inodes=478953,mode=755)
securityfs on /sys/kernel/security type securityfs (rw,nosuid,nodev,noexec,relatime)
tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev,seclabel)
devpts on /dev/pts type devpts (rw,nosuid,noexec,relatime,seclabel,gid=5,mode=620,ptmxmode=000)
tmpfs on /run type tmpfs (rw,nosuid,nodev,seclabel,mode=755)
tmpfs on /sys/fs/cgroup type tmpfs (ro,nosuid,nodev,noexec,seclabel,mode=755)
cgroup on /sys/fs/cgroup/systemd type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,xattr,release_agent=/usr/lib/systemd/systemd-cgroups-agent,name=systemd)
pstore on /sys/fs/pstore type pstore (rw,nosuid,nodev,noexec,relatime)
cgroup on /sys/fs/cgroup/Cpuset type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,cpuset)
cgroup on /sys/fs/cgroup/net_cls,net_prio type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,net_prio,net_cls)
cgroup on /sys/fs/cgroup/cpu,cpuacct type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,cpuacct,cpu)
cgroup on /sys/fs/cgroup/pids type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,PIDs)
cgroup on /sys/fs/cgroup/hugetlb type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,hugetlb)
cgroup on /sys/fs/cgroup/devices type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,devices)
cgroup on /sys/fs/cgroup/blkio type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,blkio)
cgroup on /sys/fs/cgroup/memory type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,memory)
cgroup on /sys/fs/cgroup/freezer type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,freezer)
cgroup on /sys/fs/cgroup/perf_event type cgroup (rw,nosuid,nodev,noexec,relatime,seclabel,perf_event)
configfs on /sys/kernel/config type configfs (rw,relatime)
/dev/sda1 on / type xfs (rw,relatime,seclabel,attr2,inode64,noquota)
rpc_pipefs on /var/lib/nfs/rpc_pipefs type rpc_pipefs (rw,relatime)
selinuxfs on /sys/fs/selinux type selinuxfs (rw,relatime)
debugfs on /sys/kernel/debug type debugfs (rw,relatime)
hugetlbfs on /dev/hugepages type hugetlbfs (rw,relatime,seclabel)
mqueue on /dev/mqueue type mqueue (rw,relatime,seclabel)
/dev/sdb on /data type ext4 (rw,noatime,seclabel,discard,nobarrier,data=ordered)
tmpfs on /run/user/1000 type tmpfs (rw,nosuid,nodev,relatime,seclabel,size=388020k,mode=700,uid=1000,gid=1001)
systemd-1 on /proc/sys/fs/binfmt_misc type autofs (rw,relatime,fd=43,pgrp=1,timeout=0,minproto=5,maxproto=5,direct,pipe_ino=33198)
binfmt_misc on /proc/sys/fs/binfmt_misc type binfmt_misc (rw,relatime)
tmpfs on /run/user/0 type tmpfs (rw,nosuid,nodev,relatime,seclabel,size=388020k,mode=700)
free命令可以查看buff/cache占用内存的大小,使用-w参数可以将两者分开展示,如下所示
~# free -w
total used free shared buffers cache available
Mem: 3880176 336416 748160 41480 47996 2747604 3213132
Swap: 0
free命令的输出结果中buffers表示的是buff cache大小,cache表示的page cache大小。
toptop命令中也可以展示buff/cache的大小,但是它们是被放在一起展示的,就像free命令不加-w一样,如下所示:
top - 22:40:39 up 87 days, 1:34, 2 users, load average: 0.00, 0.01, 0.05
Tasks: 94 total, 1 running, 93 sleeping, 0 stopped, 0 zombie
%Cpu(s): 0.0 us, 0.0 sy, 0.0 ni,100.0 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem : 3880176 total, 747648 free, 336800 used, 2795728 buff/cache
KiB Swap: 0 total, 0 free, 0 used. 3212748 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME COMMAND
602 root 20 0 44168 2012 1668 S 0.3 0.1 87:14.45 qemu-ga
1 root 20 0 54432 6656 4020 S 0.0 0.2 30:02.09 systemd
vmstat命令也会展示buff/cache的大小,如下所示:
~# vmstat 1
procs -----------memory---------- ---swap-- -----io---- -system-- ------cpu-----
r b swpd free buff cache si so bi bo in cs us sy id wa st
1 0 0 748052 47996 2747748 0 0 0 1 1 1 0 0 100 0 0
0 0 0 748052 47996 2747748 0 0 0 0 104 100 0 0 100 0 0
0 0 0 748036 47996 2747764 0 0 0 0 147 131 0 0 100 0 0
0 0 0 748052 47996 2747764 0 0 0 0 87 87 0 0 100 0 0
sar工具提供了多种文件系统统计方式。可以使用-v参数查看文件系统的当前活动:
~# sar -v 1
Linux 3.10.0-1160.11.1.el7.x86_64 (xxxxxxx) 07/15/2022 _x86_64_ (2 CPU)
10:44:52 PM dentunusd file-nr inode-nr pty-nr
10:44:53 PM 61343 992 56064 1
10:44:54 PM 61343 992 56064 1
10:44:55 PM 61343 992 56064 1
10:44:56 PM 61343 992 56064 1
- dentunusd:目录项缓存未使用计数(当前可用目缓存目录数)
- file-nr:打开的文件数
- inode-nr:使用的inode数
- pty-nr:使用伪终端数量
还有一个-r参数,可以展示kbbuffers和kbcached列。详情可见内存分系工具的介绍。
~# sar -r 1
Linux 3.10.0-1160.11.1.el7.x86_64 (xxxxxxxx) 07/15/2022 _x86_64_ (2 CPU)
10:48:50 PM kbmemfree kbavail kbmemused %memused kbbuffers kbcached kbcommit %commit kbactive kbinact kbdirty
10:48:51 PM 748672 3214008 225316 5.81 47996 2568488 272064 7.01 1338660 1319316 0
10:48:52 PM 748648 3213984 225340 5.81 47996 2568488 272064 7.01 1338748 1319312 0
slabtop打印了内核slab缓存,这些缓存有可能是被文件系统缓存使用的:
~# slabtop -o
Active / Total Objects (% used) : 1672622 / 1751976 (95.5%)
Active / Total Slabs (% used) : 43270 / 43270 (100.0%)
Active / Total Caches (% used) : 82 / 115 (71.3%)
Active / Total Size (% used) : 265913.80K / 287453.69K (92.5%)
Minimum / Average / Maximum Object : 0.01K / 0.16K / 15.25K
OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME
552279 540917 97% 0.10K 14161 39 56644K buffer_head
172482 162916 94% 0.04K 1691 102 6764K selinux_inode_security
136576 131676 96% 0.03K 1067 128 4268K kmalloc-32
136510 136510 100% 0.02K 803 170 3212K fsnotify_mark_connector
122560 114275 93% 0.06K 1915 64 7660K kmalloc-64
113408 108135 95% 0.02K 443 256 1772K kmalloc-16
90300 79993 88% 0.19K 4300 21 17200K dentry
84480 84480 100% 0.01K 165 512 660K kmalloc-8
75633 72534 95% 0.94K 4449 17 71184K xfs_inode
72266 68579 94% 0.17K 3142 23 12568K xfs_ili
43232 38448 88% 1.00K 2702 16 43232K kmalloc-1024
关心的内容:
- buffer_head:缓冲区告诉缓存使用项
- dentry:目录项缓存
- inode_cache:inode缓存
- ext3_inode_cache:ext3的inode缓存
- ext4_inode_cache:ext4的inode缓存
- xfs_inode:xfs的inode缓存
- btrfs_inode:btrfs的inode缓存
文件系统延迟在系统调用接口层面使用strace工具。strace当前是基于ptrace实现的,会严重影响性能,使用时要注意。如下所示,使用strace追踪crond进程查看其读取文件的延时。
~# strace -ttT -p 21621
strace: Process 21621 attached
[...]
23:04:01.944437 stat("/etc/localtime", {st_mode=S_IFREG|0644, st_size=556, ...}) = 0 <0.000026>
[...]
<detached ...>
perf也是一个强大的工具,可以追中文件系统的tracepoint,使用kprobe查看VFS和文件系统的内部,并且还有一个trace子命令比strace高效的多。
~# perf trace cksum /usr/bin/cksum
759126066 33152 /usr/bin/cksum
0.036 ( 0.002 ms): cksum/9147 brk() = 0x1d13000
0.078 ( 0.005 ms): cksum/9147 mmap(len: 4096, prot: READ|WRITE, flags: PRIVATE|ANONYMOUS) = 0x7f42421d5000
0.096 ( 0.008 ms): cksum/9147 access(filename: 0x41fd3cb0, mode: R) = -1 ENOENT (No such file or directory)
0.117 ( 0.006 ms): cksum/9147 open(filename: 0x41fd25a4, flags: RDONLY|CLOEXEC) = 3
0.125 ( 0.003 ms): cksum/9147 fstat(fd: 3, statbuf: 0x7ffd11e0bb70) = 0
0.129 ( 0.004 ms): cksum/9147 mmap(len: 35679, prot: READ, flags: PRIVATE, fd: 3)
可以使用perf list命令列出感兴趣的tracepiont或者kprobe:
~# perf list *xfs*
List of pre-defined events (to be used in -e):
xfs:xfs_agf [Tracepoint event]
xfs:xfs_agfl_reset [Tracepoint event]
xfs:xfs_ail_delete [Tracepoint event]
xfs:xfs_ail_flushing [Tracepoint event]
xfs:xfs_ail_insert [Tracepoint event]
xfs:xfs_ail_locked [Tracepoint event]
xfs:xfs_ail_move [Tracepoint event]
xfs:xfs_ail_pinned [Tracepoint event]
可以使用perf stat统计xfs trancepoint调用,如下所示:
> perf stat -e 'xfs:*' -a
^C
Performance counter stats for 'system wide':
1 xfs:xfs_ilock
1 xfs:xfs_iunlock
1 xfs:xfs_filemap_fault
1 xfs:xfs_log_force
21.508754970 seconds time elapsed
还可以使用perf record将获取结果存如文件,使用perf script读取文件内容进行分析。
fatracefatrace使用Linux的fanotify API,示例输出:
~# fatrace
rsyslogd(968876): W /var/log/syslog
rsyslogd(968876): W /var/log/auth.log
gsd-color(1690): CO /usr/share/zoneinfo/Etc/UTC
gsd-color(1690): CO /usr/share/zoneinfo/Etc/UTC
gsd-housekeepin(1707): RO /etc/fstab
gsd-housekeepin(1707): C /etc/fstab
每一行显示了进程名、PID、事件类型、全路径和可选的状态。时间类型包含:
- O:打开
- R:读取
- W:写入
- C:关闭
例如:CW 代表关闭一个写文件,CO关闭一个打开文件
BPFopensnoop跟踪文件打开事件
BCC版本
~# opensnoop-bpfcc -T
TIME(s) PID COMM FD ERR PATH
0.000000000 1700193 hidsagent 4 0 /proc/net/dev
0.000075000 1700193 hidsagent 4 0 /proc/stat
0.000117000 1700193 hidsagent 4 0 /proc/1700193/stat
0.000148000 1700193 hidsagent 4 0 /proc/1700193/task
0.000213000 1700193 hidsagent 14 0 /proc/1700193/task/1700193/stat
0.000228000 1700193 hidsagent 14 0 /proc/1700193/task/1700194/stat
0.000242000 1700193 hidsagent 14 0 /proc/1700193/task/1700196/stat
BCC命令用法
opensnoop [options]
命令选项
- -x:只显示打开失败的操作
- -p PID:仅监控给定的进程
- -n NAME:仅显示进程名字包含NAME的事件
bpftrace版本不支持任何参数,其源码如下所示:
BEGIN
{
printf("Tracing open syscalls... Hit Ctrl-C to end.\n");
printf("%-6s %-16s %4s %3s %s\n", "PID", "COMM", "FD", "ERR", "PATH");
}
tracepoint:syscalls:sys_enter_open,
tracepoint:syscalls:sys_enter_openat
{
@filename[tid] = args->filename;
}
tracepoint:syscalls:sys_exit_open,
tracepoint:syscalls:sys_exit_openat
/@filename[tid]/
{
$ret = args->ret;
$fd = $ret > 0 ? $ret : -1;
$errno = $ret > 0 ? 0 : - $ret;
printf("%-6d %-16s M = %s\n", pid, comm, $fd, $errno,
str(@filename[tid]));
delete(@filename[tid]);
}
END
{
clear(@filename);
}
执行结果示例:
~# opensnoop.bt
Attaching 6 probes...
Tracing open syscalls... Hit Ctrl-C to end.
PID COMM FD ERR PATH
884 snmpd 2 0 /proc/net/dev
884 snmpd 2 0 /proc/net/if_inet6
884 snmpd 2 0 /sys/class/net/eth0/device/vendor
884 snmpd 2 0 /sys/class/net/eth0/device/device
针对stats系统调用
BCC版本
~# statsnoop-bpfcc
PID COMM FD ERR PATH
2582812 zabbix_agentd 0 0 /var/log/zabbix/zabbix_agentd.log
2582812 zabbix_agentd 0 0 /etc/resolv.conf
1700214 hidsagent 0 0 /home/sfop/security/hids/log/net_monitor.log
1700214 hidsagent 0 0 /home/sfop/security/hids/log/net_monitor.log
1700214 hidsagent 0 0 /home/sfop/security/hids/log/net_monitor.log
1700214 hidsagent 0 0 /home/sfop/security/hids/log/net_monitor.log
2582816 zabbix_agentd 0 0 /var/log/zabbix/zabbix_agentd.log
BCC命令行用法
statsnoop [options]
命令行选项
- -x:仅显示失败的stat调用
- -t:增加一列时间戳信息(秒)
- -p PID:仅测量给定的PID
该版本不支持任何选项,使用示例如下:
~# statsnoop.bt
Attaching 10 probes...
Tracing stat syscalls... Hit Ctrl-C to end.
PID COMM ERR PATH
925 cron 0 /etc/localtime
925 cron 0 crontabs
925 cron 0 /etc/crontab
925 cron 0 /etc/cron.d
925 cron 0 /etc/cron.d/php
925 cron 0 /etc/cron.d/popularity-contest
925 cron 0 /etc/cron.d/e2scrub_all
925 cron 0 /etc/cron.d/sysstat
1707 gsd-housekeepin 0 /etc/fstab
1707 gsd-housekeepin 0 /proc
1707 gsd-housekeepin 0 /proc/self