gluster增加xlator操作（一）

查看更多，请猛点向glusterfs中增加xlator（一）
前面已经介绍过，glusterfs中主线为xlator，几乎每一个大的功能点，或者性能，都可以以xlator形式参与进glusterfs中，那么xlator的开发就必不可少，所以，这里介绍一下简单的向glusterfs中增加xlator的方法，首先创建一个卷testvol

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17

[root@CM addxlator]# mkdir /opt/test [root@CM addxlator]# gluster volume create testvol 192.168.0.194:/opt/test force volume create: testvol: success: please start the volume to access data [root@CM addxlator]# gluster volume info all    Volume Name: testvol Type: Distribute Volume ID: f0ef2a09-09c0-4a84-aeb6-3f2e0aebf4c2 Status: Created Number of Bricks: 1 Transport-type: tcp Bricks: Brick1: 192.168.0.194:/opt/test [root@CM addxlator]# [root@CM addxlator]# gluster volume start testvol force volume start: testvol: success [root@CM addxlator]#

创建成功后，可以查看volume file中的内容

 cat /var/lib/glusterd/vols/testvol/testvol-fuse.vol 

01 02 03 04 05 06 07 08 09 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

volume testvol-client-0     type protocol/client     option transport-type tcp     option remote-subvolume /opt/test     option remote-host 192.168.0.194 end-volume   volume testvol-dht     type cluster/distribute     subvolumes testvol-client-0 end-volume   volume testvol-write-behind     type performance/write-behind     subvolumes testvol-dht end-volume   volume testvol-read-ahead     type performance/read-ahead     subvolumes testvol-write-behind end-volume   volume testvol-io-cache     type performance/io-cache     subvolumes testvol-read-ahead end-volume   volume testvol-quick-read     type performance/quick-read     subvolumes testvol-io-cache end-volume   volume testvol-open-behind     type performance/open-behind     subvolumes testvol-quick-read end-volume   volume testvol-md-cache     type performance/md-cache     subvolumes testvol-open-behind end-volume   volume testvol     type debug/io-stats     option count-fop-hits off     option latency-measurement off     subvolumes testvol-md-cache end-volume

这个配置为默认的配置，接下来需要将自己的xlator增加入该配置中，根据之前所提到过的xlator_dyload的过程，需要修改这个volume file，来增加xlator，修改配置如下：

01 02 03 04 05 06 07 08 09 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

volume testvol-client-0     type protocol/client     option transport-type tcp     option remote-subvolume /opt/test     option remote-host 192.168.0.194 end-volume   volume testvol-dht     type cluster/distribute     subvolumes testvol-client-0 end-volume   volume testvol-write-behind     type performance/write-behind     subvolumes testvol-dht end-volume   volume testvol-read-ahead     type performance/read-ahead     subvolumes testvol-write-behind end-volume   volume testvol-io-cache     type performance/io-cache     subvolumes testvol-read-ahead end-volume   volume testvol-quick-read     type performance/quick-read     subvolumes testvol-io-cache end-volume   volume testvol-open-behind     type performance/open-behind     subvolumes testvol-quick-read end-volume   volume testvol-md-cache     type performance/md-cache     subvolumes testvol-open-behind end-volume volume testvol-test     type debug/test     subvolumes testvol-md-cache end-volume volume testvol     type debug/io-stats     option count-fop-hits off     option latency-measurement off     subvolumes testvol-test end-volume

可以看到，配置中多了个testvol-test这么一层，接下来将glusterfs挂载至挂载点查看输出log

1 2 3

[root@CM addxlator]# [root@CM addxlator]# mount -t glusterfs 192.168.0.194:testvol /media [root@CM addxlator]#

查看gluster挂载到的挂载点对应的log输出信息
cat /usr/local/var/log/glusterfs/media.log

01 02 03 04 05 06 07 08 09 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

[2013-06-10 11:00:29.899440] I [glusterfsd.c:1878:main] 0-/usr/local/sbin/glusterfs: Started running /usr/local/sbin/glusterfs version 3git (/usr/local/sbin/glusterfs --volfile-id=testvol --volfile-server=192.168.0.194 /media) [2013-06-10 11:00:29.903627] I [socket.c:3515:socket_init] 0-glusterfs: SSL support is NOT enabled [2013-06-10 11:00:29.903823] I [socket.c:3530:socket_init] 0-glusterfs: using system polling thread [2013-06-10 11:00:29.909423] E [test.c:54:init] 0-testvol-test: test translator loaded [2013-06-10 11:00:29.909718] I [dht-shared.c:311:dht_init_regex] 0-testvol-dht: using regex rsync-hash-regex = ^\.(.+)\.[^.]+$ [2013-06-10 11:00:29.911047] I [socket.c:3515:socket_init] 0-testvol-client-0: SSL support is NOT enabled [2013-06-10 11:00:29.911173] I [socket.c:3530:socket_init] 0-testvol-client-0: using system polling thread [2013-06-10 11:00:29.911218] I [client.c:2236:notify] 0-testvol-client-0: parent translators are ready, attempting connect on transport Final graph: +------------------------------------------------------------------------------+   1: volume testvol-client-0   2:     type protocol/client   3:     option remote-host 192.168.0.194   4:     option remote-subvolume /opt/test   5:     option transport-type socket   6: end-volume   7:   8: volume testvol-dht   9:     type cluster/distribute  10:     subvolumes testvol-client-0  11: end-volume  12:  13: volume testvol-write-behind  14:     type performance/write-behind  15:     subvolumes testvol-dht  16: end-volume  17:  18: volume testvol-read-ahead  19:     type performance/read-ahead  20:     subvolumes testvol-write-behind  21: end-volume  22:  23: volume testvol-io-cache  24:     type performance/io-cache  25:     subvolumes testvol-read-ahead  26: end-volume  27:  28: volume testvol-quick-read  29:     type performance/quick-read  30:     subvolumes testvol-io-cache  31: end-volume  32:  33: volume testvol-open-behind  34:     type performance/open-behind  35:     subvolumes testvol-quick-read  36: end-volume  37:  38: volume testvol-md-cache  39:     type performance/md-cache  40:     subvolumes testvol-open-behind  41: end-volume  42:  43: volume testvol-test  44:     type debug/test  45:     subvolumes testvol-md-cache  46: end-volume  47:  48: volume testvol  49:     type debug/io-stats  50:     option latency-measurement off  51:     option count-fop-hits off  52:     subvolumes testvol-test  53: end-volume  54: +------------------------------------------------------------------------------+ [2013-06-10 11:00:29.914847] I [rpc-clnt.c:1675:rpc_clnt_reconfig] 0-testvol-client-0: changing port to 49152 (from 0) [2013-06-10 11:00:29.918227] I [client-handshake.c:1658:select_server_supported_programs] 0-testvol-client-0: Using Program GlusterFS 3.3, Num (1298437), Version (330) [2013-06-10 11:00:29.918482] I [client-handshake.c:1456:client_setvolume_cbk] 0-testvol-client-0: Connected to 192.168.0.194:49152, attached to remote volume '/opt/test'. [2013-06-10 11:00:29.918548] I [client-handshake.c:1468:client_setvolume_cbk] 0-testvol-client-0: Server and Client lk-version numbers are not same, reopening the fds [2013-06-10 11:00:29.923683] I [fuse-bridge.c:4767:fuse_graph_setup] 0-fuse: switched to graph 0 [2013-06-10 11:00:29.923986] I [client-handshake.c:450:client_set_lk_version_cbk] 0-testvol-client-0: Server lk version = 1 [2013-06-10 11:00:29.924164] I [fuse-bridge.c:3724:fuse_init] 0-glusterfs-fuse: FUSE inited with protocol versions: glusterfs 7.13 kernel 7.13 [2013-06-10 11:00:29.924225] E [test.c:28:test_lookup] 0-testvol-test: in test translator lookup [2013-06-10 11:00:29.925153] E [test.c:28:test_lookup] 0-testvol-test: in test translator lookup

经过查看log，可以看到test模块中的打印输出， in test translator lookup
接下来查看描述一下外挂xlator的方式
首先现在glusterfs的源代码，编译以后，在本地也就存在了glusterfs的开发环境，为了开发调试方便，可暂不将xlator存于glusterfs工程中，这时可以编写一个Makefile
首先建立一个目录

1 2	mkdir addxlator_sample cd  addxlator_sample

Makefile中内容可如下：

01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

#Add translator into glusterfs volume file #Author Steven Liu #E-mail [e-mail] #Blog: http://blog.fs-linux.org   TARGET  = test.so OBJECTS = test.o   GLUSTERFS_SRC   = /opt/glusterfs GLUSTERFS_LIB   = /usr/local/lib HOST_OS = HF_LINUX_HOST_OS   CFLAGS  = -fPIC -Wall -O0 -g \           -DHAVE_CONFIG_H -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE -D$(HOST_OS) \           -I$(GLUSTERFS_SRC) -I$(GLUSTERFS_SRC)/libglusterfs/src \           -I$(GLUSTERFS_SRC)/contrib/uuid   LDFLAGS = -shared -nostartfiles -L$(GLUSTERFS_LIB) -lglusterfs -lpthread   $(TARGET): $(OBJECTS)     $(CC) $(OBJECTS) $(LDFLAGS) -o $(TARGET)     clean:     rm -rf $(TARGET) $(OBJECTS)

Makefile编写完成之后，需要OBJECTS，这时需要编写test.c来保证编译通过，当然，test这个xlator也需要在这里生成

001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 029 030 031 032 033 034 035 036 037 038 039 040 041 042 043 044 045 046 047 048 049 050 051 052 053 054 055 056 057 058 059 060 061 062 063 064 065 066 067 068 069 070 071 072 073 074 075 076 077 078 079 080 081 082 083 084 085 086 087 088 089 090 091 092 093 094 095 096 097 098 099 100 101 102 103 104 105 106 107 108 109 110 111

#ifndef _CONFIG_H #define _CONFIG_H #include "config.h" #include "xlator.h" #endif   #include #include #include "glusterfs.h" #include "xlator.h" #include #include "defaults.h" #include "logging.h"   int test_lookup_cbk (call_frame_t *frame, void *cookie, xlator_t *this,                      int32_t op_ret, int32_t op_errno,                      inode_t *inode, struct iatt *buf,                      dict_t *xdata, struct iatt *postparent) {         STACK_UNWIND_STRICT (lookup, frame, op_ret, op_errno, inode, buf, xdata,                              postparent);         return 0; }     static int test_lookup(call_frame_t *frame, xlator_t *this, loc_t *loc, dict_t *xdata) {     gf_log(this->name, GF_LOG_ERROR, "in test translator lookup");     STACK_WIND (frame, test_lookup_cbk,             FIRST_CHILD(this), FIRST_CHILD(this)->fops->lookup,             loc, xdata);     return 0; }   static int test_stat (call_frame_t *frame, xlator_t *this, loc_t *loc, dict_t *xdata) {     gf_log(this->name, GF_LOG_ERROR, "in test translator stat");     return 0; }   int reconfigure (xlator_t *this, dict_t *options) {         return 0; }     int init (xlator_t *this) {         struct ios_conf    *conf = NULL;         int                 ret = -1;       gf_log (this->name, GF_LOG_ERROR, "test translator loaded");         if (!this)                 return -1;           if (!this->children) {                 gf_log (this->name, GF_LOG_ERROR,                         "test translator requires atleast one subvolume");                 return -1;         }           if (!this->parents) {               gf_log (this->name, GF_LOG_ERROR, "dangling volume. check volfile ");         }           conf = this->private;           this->private = conf;         ret = 0;         return ret; }     void fini (xlator_t *this) {         struct ios_conf *conf = NULL;           if (!this)                 return;           conf = this->private;           if (!conf)                 return;         this->private = NULL;           GF_FREE(conf);         gf_log (this->name, GF_LOG_ERROR, "test translator unloaded");         return; }   int notify (xlator_t *this, int32_t event, void *data, ...) {     default_notify (this, event, data);         return 0; }   struct xlator_fops fops = {         .stat        = test_stat,         .lookup      = test_lookup, };   struct xlator_cbks cbks = { };   struct volume_options options[] = {  };

目前仅用于测试的xlator而已，所以自己定义了lookup与stat，还有lookup_cbk，其他的并未定义，全部走default，这个原理在之前已经提到过。
然后执行make既可生成对应的xlator文件

1	Makefile  test.c  test.o  test.so

将test.so复制至/usr/local/lib/glusterfs/3git/xlator/debug中
然后执行glusterfs的客户端，mount至挂载节点，如果成功，可以查看log，整个过程成功矣~~！