客户端挂载和管理glusterfs卷
目录
一、客户端挂载
可以使用Gluster Native Client方法在GNU / Linux客户端中实现高并发性，性能和透明故障转移。可以使用NFS v3访问gluster卷。已经对GNU / Linux客户端和其他操作系统中的NFS实现进行了广泛的测试，例如FreeBSD，Mac OS X，以及Windows 7（Professional和Up）和Windows Server 2003.其他NFS客户端实现可以与gluster一起使用NFS服务器。使用Microsoft Windows以及SAMBA客户端时，可以使用CIFS访问卷。对于此访问方法，Samba包需要存在于客户端。
总结：GlusterFS支持三种客户端类型。Gluster Native Client、NFS和CIFS。Gluster Native Client是在用户空间中运行的基于FUSE的客户端，官方推荐使用Native Client，可以使用GlusterFS的全部功能。
1.1 使用Gluster Native Client挂载
Gluster Native Client是基于FUSE的，所以需要保证客户端安装了FUSE。这个是官方推荐的客户端，支持高并发和高效的写性能。
在开始安装Gluster Native Client之前，您需要验证客户端上是否已加载FUSE模块，并且可以访问所需的模块，如下所示：
$ modprobe fuse #将FUSE可加载内核模块（LKM）添加到Linux内核
$ dmesg | grep -i fuse #验证是否已加载FUSE模块
[ 569.630373] fuse init (API version 7.22)
1.2 安装Gluseter Native Client
$ yum -y install glusterfs-client #安装glusterfs-client客户端
$ mkdir /mnt/glusterfs #创建挂载目录
$ mount.glusterfs 192.168.56.11:/gv1 /mnt/glusterfs/ #挂载/gv1
$ df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda2 20G 1.4G 19G 7% /
devtmpfs 231M 0 231M 0% /dev
tmpfs 241M 0 241M 0% /dev/shm
tmpfs 241M 4.6M 236M 2% /run
tmpfs 241M 0 241M 0% /sys/fs/cgroup
/dev/sda1 197M 97M 100M 50% /boot
tmpfs 49M 0 49M 0% /run/user/0
192.168.56.11:/gv1 4.0G 312M 3.7G 8% /mnt/glusterfs
$ ll /mnt/glusterfs/ #查看挂载目录的内容
total 100000
-rw-r--r-- 1 root root 102400000 Aug 7 04:30 100M.file
$ mount #查看挂载信息
sysfs on /sys type sysfs (rw,nosuid,nodev,noexec,relatime)
proc on /proc type proc (rw,nosuid,nodev,noexec,relatime)
......
192.168.99.251:/gv1 on /mnt/glusterfs type fuse.glusterfs (rw,relatime,user_id=0,group_id=0,default_permissions,allow_other,max_read=131072)
1.3 手动挂载卷选项
使用该mount -t glusterfs命令时，可以指定以下选项 。请注意，您需要用逗号分隔所有选项。
backupvolfile-server=server-name #在安装fuse客户端时添加了这个选择，则当第一个vofile服务器故障时，该选项执行的的服务器将用作volfile服务器来安装客户端
volfile-max-fetch-attempts=number of attempts 指定在装入卷时尝试获取卷文件的尝试次数。
log-level=loglevel #日志级别
log-file=logfile #日志文件
transport=transport-type #指定传输协议
direct-io-mode=[enable|disable]
use-readdirp=[yes|no] #设置为ON，则强制在fuse内核模块中使用readdirp模式
举个例子：
# mount -t glusterfs -o backupvolfile-server=volfile_server2,use-readdirp=no,volfile-max-fetch-attempts=2,log-level=WARNING,log-file=/var/log/gluster.log server1:/test-volume /mnt/glusterfs
1.4 自动挂载卷
除了使用mount挂载，还可以使用/etc/fstab自动挂载
语法格式：HOSTNAME-OR-IPADDRESS:/VOLNAME MOUNTDIR glusterfs defaults,_netdev 0 0
举个例子：
192.168.99.251:/gv1 /mnt/glusterfs glusterfs defaults,_netdev 0 0
二、管理GlusterFS卷
2.1 停止卷
$ gluster volume stop gv1
2.2 删除卷
$ gluster volume delete gv1
2.3 扩展卷
GlusterFS支持在线进行卷的扩展。
如果添加的节点还不是集群中的节点，需要使用下面命令添加到集群
语法：gluster peer probe <SERVERNAME>
扩展卷语法：gluster volume add-brick <VOLNAME> <NEW-BRICK>
$ gluster peer probe gluster-node3 #添加gluster-node3到集群
peer probe: success.
$ gluster volume add-brick test-volume gluster-node3:/storage/brick1 force #扩展test-volume卷
volume add-brick: success
$ gluster volume info
Volume Name: test-volume
Type: Distribute
Volume ID: 26a625bb-301c-4730-a382-0a838ee63935
Status: Started
Snapshot Count: 0
Number of Bricks: 3
Transport-type: tcp
Bricks:
Brick1: gluster-node1:/storage/brick1
Brick2: gluster-node2:/storage/brick1
Brick3: gluster-node3:/storage/brick1 #增加的brick
Options Reconfigured:
transport.address-family: inet
nfs.disable: on
$ gluster volume rebalance test-volume start #添加后，重新平衡卷以确保文件分发到新添加的brick
volume rebalance: test-volume: success: Rebalance on test-volume has been started successfully. Use rebalance status command to check status of the rebalance process.
ID: ca58bd21-11a5-4018-bb2a-8f9079982394
2.4 收缩卷
收缩卷和扩展卷相似据以Brick为单位。
语法：gluster volume remove-brick <VOLNAME> <BRICKNAME> start
$ gluster volume remove-brick test-volume gluster-node3:/storage/brick1 start #删除brick
volume remove-brick start: success
ID: dd0004f0-b3e6-45d6-80ed-90506dc16159
$ gluster volume remove-brick test-volume gluster-node3:/storage/brick1 status #查看remove brick操作的状态
Node Rebalanced-files size scanned failures skipped status run time in h:m:s
--------- ----------- ----------- ----------- ----------- ----------- ------------ --------------
gluster-node3 35 0Bytes 35 0 0 completed 0:00:00
$ gluster volume remove-brick test-volume gluster-node3:/storage/brick1 commit #显示completed状态后，提交remove-brick操作
volume remove-brick commit: success
$ gluster volume info
Volume Name: test-volume
Type: Distribute
Volume ID: 26a625bb-301c-4730-a382-0a838ee63935
Status: Started
Snapshot Count: 0
Number of Bricks: 2
Transport-type: tcp
Bricks:
Brick1: gluster-node1:/storage/brick1
Brick2: gluster-node2:/storage/brick1
Options Reconfigured:
performance.client-io-threads: on
transport.address-family: inet
nfs.disable: on
2.5 迁移卷
要替换分布式卷上的brick，需要添加一个新的brick，然后删除要替换的brick。在替换的过程中会触发重新平衡的操作，会将移除的brick中的数据到新加入的brick中。
注意：这里仅支持可以对分布式复制卷或复制卷使用"replace-brick"命令进行替换操作。
（1）初始卷test-volume的配置信息$ gluster volume info
Volume Name: test-volume
Type: Distribute
Volume ID: 26a625bb-301c-4730-a382-0a838ee63935
Status: Started
Snapshot Count: 0
Number of Bricks: 2
Transport-type: tcp
Bricks:
Brick1: gluster-node1:/storage/brick1
Brick2: gluster-node2:/storage/brick1
Options Reconfigured:
performance.client-io-threads: on
transport.address-family: inet
nfs.disable: on
（2）test-volume挂载目录的文件和在实际存储位置的文件信息
$ ll
total 0
-rw-r--r-- 1 root root 0 Aug 13 22:22 file1
-rw-r--r-- 1 root root 0 Aug 13 22:22 file2
-rw-r--r-- 1 root root 0 Aug 13 22:22 file3
-rw-r--r-- 1 root root 0 Aug 13 22:22 file4
-rw-r--r-- 1 root root 0 Aug 13 22:22 file5
$ ll /storage/brick1/
total 0
-rw-r--r-- 2 root root 0 Aug 13 22:22 file1
-rw-r--r-- 2 root root 0 Aug 13 22:22 file2
-rw-r--r-- 2 root root 0 Aug 13 22:22 file5
$ ll /storage/brick1/
total 0
-rw-r--r-- 2 root root 0 Aug 13 2018 file3
-rw-r--r-- 2 root root 0 Aug 13 2018 file4（3）添加新brick gluster-node3:/storage/brick1
$ gluster volume add-brick test-volume gluster-node3:/storage/brick1/ force
volume add-brick: success
（4）启动remove-brick
$ gluster volume remove-brick test-volume gluster-node2:/storage/brick1 start
volume remove-brick start: success
ID: 2acdaebb-25a9-477c-807e-980a6086796e
（5）查看remove-brick的状态是否为completed
$ gluster volume remove-brick test-volume gluster-node2:/storage/brick1 status
Node Rebalanced-files size scanned failures skipped status run time in h:m:s
--------- ----------- ----------- ----------- ----------- ----------- ------------ --------------
gluster-node2 2 0Bytes 2 0 0 completed 0:00:00
（6）确认删除旧的brick
$ gluster volume remove-brick test-volume gluster-node2:/storage/brick1 commit
volume remove-brick commit: success
（7）test-volume的最新配置
$ gluster volume info
Volume Name: test-volume
Type: Distribute
Volume ID: 26a625bb-301c-4730-a382-0a838ee63935
Status: Started
Snapshot Count: 0
Number of Bricks: 2
Transport-type: tcp
Bricks:
Brick1: gluster-node1:/storage/brick1
Brick2: gluster-node3:/storage/brick1
Options Reconfigured:
performance.client-io-threads: on
transport.address-family: inet
nfs.disable: on
（8）检查新增brick的文件存储信息，原先存储在gluster-node2节点的文件移动到了gluster-node3中
[root@gluster-node3 ~]# ll /storage/brick1/
total 0
-rw-r--r-- 2 root root 0 Aug 13 2018 file3
-rw-r--r-- 2 root root 0 Aug 13 2018 file4
2.6 系统配额
$ gluster volume quota test-volume enable #启用配额
volume quota : success
$ gluster volume quota test-volume disable #禁用配额
volume quota : success
$ mount -t glusterfs 127.0.0.1:/test-volume /gv1 #挂载test-volume卷
$ mkdir /gv1/quota #创建限制的目录
$ gluster volume quota test-volume limit-usage /quota 10MB #对/gv1/quota目录限制
$ gluster volume quota test-volume list #查看目录限制信息
Path Hard-limit Soft-limit Used Available Soft-limit exceeded? Hard-limit exceeded?
-------------------------------------------------------------------------------------------------------------------------------
/quota 10.0MB 80%(8.0MB) 0Bytes 10.0MB No No
$ gluster volume set test-volume features.quota-timeout 5 #设置信息的超时时间
[root@gluster-node1 quota]# cp /gv1/20M.file . #拷贝20M文件到/gv1/quota下，已经超出了限额，但是依旧可以成功，由于限制的值较小，可能受到算法的影响
[root@gluster-node1 quota]# cp /gv1/20M.file ./20Mb.file #再拷贝20M的文件，就会提示超出目录限额
cp: cannot create regular file ‘./20Mb.file’: Disk quota exceeded
$ gluster volume quota test-volume remove /quota #删除某个目录的quota设置
volume quota : success
备注：
quota功能，主要是对挂载点下的某个目录进行空间限额，如：/mnt/glusterfs/data目录，而不是对组成卷组的空间进行限制。
2.7 I/O信息查看
Profile Command 提供接口查看一个卷中的每一个brick的IO信息。
$ gluster volume profile test-volume start #启动profiling，之后则可以进行IO信息查看
Starting volume profile on test-volume has been successful
$ gluster volume profile test-volume info #查看IO信息，可以查看到每个brick的IO信息
Brick: gluster-node1:/storage/brick1
------------------------------------
Cumulative Stats:
Block Size: 32768b+ 131072b+
No. of Reads: 0 0
No. of Writes: 2 312
%-latency Avg-latency Min-Latency Max-Latency No. of calls Fop
--------- ----------- ----------- ----------- ------------ ----
0.00 0.00 us 0.00 us 0.00 us 122 FORGET
0.00 0.00 us 0.00 us 0.00 us 160 RELEASE
0.00 0.00 us 0.00 us 0.00 us 68 RELEASEDIR
Duration: 250518 seconds
Data Read: 0 bytes
Data Written: 40960000 bytes
Interval 1 Stats:
Duration: 27 seconds
Data Read: 0 bytes
Data Written: 0 bytes
Brick: gluster-node3:/storage/brick1
------------------------------------
Cumulative Stats:
Block Size: 1024b+ 2048b+ 4096b+
No. of Reads: 0 0 0
No. of Writes: 3 1 10
Block Size: 8192b+ 16384b+ 32768b+
No. of Reads: 0 0 1
No. of Writes: 291 516 68
Block Size: 65536b+ 131072b+
No. of Reads: 0 156
No. of Writes: 6 20
%-latency Avg-latency Min-Latency Max-Latency No. of calls Fop
--------- ----------- ----------- ----------- ------------ ----
0.00 0.00 us 0.00 us 0.00 us 3 RELEASE
0.00 0.00 us 0.00 us 0.00 us 31 RELEASEDIR
Duration: 76999 seconds
Data Read: 20480000 bytes
Data Written: 20480000 bytes
Interval 1 Stats:
Duration: 26 seconds
Data Read: 0 bytes
Data Written: 0 bytes
$ gluster volume profile test-volume stop #查看结束后关闭profiling功能
Stopping volume profile on test-volume has been successful
2.8 Top监控
Top command 允许你查看bricks的性能例如：read, write, file open calls, file read calls, file write calls, directory open calls, and directory real calls
所有的查看都可以设置top数，默认100
# gluster volume top VOLNAME open [brick BRICK-NAME] [list-cnt] //查看打开的fd
$ gluster volume top test-volume open brick gluster-node1:/storage/brick1 list-cnt 3
Brick: gluster-node1:/storage/brick1
Current open fds: 0, Max open fds: 4, Max openfd time: 2018-08-13 11:53:24.099217
Count filename
=======================
1 /98.txt
1 /95.txt
1 /87.txt
# gluster volume top VOLNAME read [brick BRICK-NAME] [list-cnt] //查看调用次数最多的读调用
$ gluster volume top test-volume read brick gluster-node3:/storage/brick1
Brick: gluster-node3:/storage/brick1
Count filename
=======================
157 /20M.file
# gluster volume top VOLNAME write [brick BRICK-NAME] [list-cnt] //查看调用次数最多的写调用
$ gluster volume top test-volume write brick gluster-node3:/storage/brick1
Brick: gluster-node3:/storage/brick1
Count filename
=======================
915 /20M.file
# gluster volume top VOLNAME opendir [brick BRICK-NAME] [list-cnt]
# gluster volume top VOLNAME readdir [brick BRICK-NAME] [list-cnt] //查看次数最多的目录调用
$ gluster volume top test-volume opendir brick gluster-node3:/storage/brick1
Brick: gluster-node3:/storage/brick1
Count filename
=======================
7 /quota
$ gluster volume top test-volume readdir brick gluster-node3:/storage/brick1
Brick: gluster-node3:/storage/brick1
Count filename
=======================
7 /quota
# gluster volume top VOLNAME read-perf [bsblk-size count count] [brick BRICK-NAME] [list-cnt] //查看每个Brick的读性能
$ gluster volume top test-volume read-perf bs 256 count 1 brick gluster-node3:/storage/brick1
Brick: gluster-node3:/storage/brick1
Throughput 42.67 MBps time 0.0000 secs
MBps Filename Time
==== ======== ====
0 /20M.file 2018-08-14 03:32:24.7443
# gluster volume top VOLNAME write-perf [bsblk-size count count] [brick BRICK-NAME] [list-cnt] //查看每个Brick的写性能
$ gluster volume top test-volume write-perf bs 256 count 1 brick gluster-node1:/storage/brick1
Brick: gluster-node1:/storage/brick1
Throughput 16.00 MBps time 0.0000 secs
MBps Filename Time
==== ======== ====
0 /quota/20Mb.file 2018-08-14 11:34:21.957635
0 /quota/20M.file 2018-08-14 11:31:02.767068
*************** 当你发现自己的才华撑不起野心时，就请安静下来学习吧！***************
