在配置完多台dataserver之后，就需要考虑到nameserver的单点故障问题。本文将介绍如何针对tfs的nameserver实现ha，官方推荐采用heartbeat实现，但heartbeat的实现配置同keepalive相比，复杂了许多，因而这里我们采用keepalive来实现namserver的ha和failover。
环境介绍：
NS服务器1 ——> 192.168.1.225
NS服务器2 ——> 192.168.1.226
NS VIP ——> 192.168.1.229
Data server服务器1 ——> 192.168.1.227
Data server服务器2 ——> 192.168.1.228
Nginx server ——> 192.168.1.12
在开始之前，先关闭tfs的所有的name server和data server
# /usr/local/tfs/scripts/tfs stop_ns nameserver exit SUCCESSFULLY # /usr/local/tfs/scripts/tfs stop_ds 1-3 dataserver 1 exit SUCCESSFULLY dataserver 2 exit SUCCESSFULLY dataserver 3 exit SUCCESSFULLY
一：配置tfs name server
225服务器： # grep -v ‘^#‘ /usr/local/tfs/conf/ns.conf |grep -v ‘^$‘ [public] log_size=1073741824 log_num = 16 log_level=info task_max_queue_size = 10240 port = 8108 work_dir=/usr/local/tfs dev_name= eth0 thread_count = 4 ip_addr = 192.168.1.229 //229地址为vip [nameserver] safe_mode_time = 300 ip_addr_list = 192.168.1.225|192.168.1.226 group_mask = 255.255.255.255 block_max_size = 83886080 max_replication = 2 min_replication = 2 use_capacity_ratio = 98 block_max_use_ratio = 98 heart_interval = 2 object_dead_max_time = 3600 cluster_id = 1 replicate_ratio_ = 50 max_write_filecount = 16 heart_thread_count = 2 heart_max_queue_size = 10 repl_max_time = 60 compact_delete_ratio = 15 compact_max_load = 200 object_dead_max_time = 86400 object_clear_max_time = 300 max_wait_write_lease = 15 lease_expired_time = 3 max_lease_timeout = 3000 cleanup_lease_threshold = 102400 build_plan_interval = 10 run_plan_expire_interval = 120 build_plan_ratio = 25 dump_stat_info_interval = 60000000 build_plan_default_wait_time = 2 balance_max_diff_block_num = 5 add_primary_block_count = 3 block_chunk_num = 32 task_percent_sec_size = 200 task_max_queue_size = 10000 oplog_sync_max_slots_num = 1024 oplog_sync_thread_num = 1 # cd /usr/local/ # tar -zcvpf tfs.tgz tfs/ # scp tfs.tgz 192.168.1.226:/usr/local/ 226服务器： # cd /usr/local/ # tar -zxvpf tfs.tgz
二：配置ha
225服务器
# cd /usr/local/src/
# wget http://www.keepalived.org/software/keepalived-1.2.13.tar.gz
# tar -zxvpf keepalived-1.2.13.tar.gz
# cd keepalived-1.2.13
# ./configure --prefix=/usr/local/keepalived
# make && make install
# cat /usr/local/keepalived/etc/keepalived/keepalived.conf
global_defs {
router_id tfs_ns
}
vrrp_script chk_nameserver {
script "killall -0 nameserver"
interval 2
weight 2
}
vrrp_instance VI_1 {
state MASTER
interface eth0
virtual_router_id 23
priority 101
advert_int 1
authentication {
auth_type PASS
auth_pass password_123
}
virtual_ipaddress {
192.168.1.229/24 brd 192.168.255.255 dev eth0 label eth0:0
}
track_interface {
eth1
}
track_script {
chk_nameserver
}
}
# cd /usr/local/
# tar -zcvpf keepalived.tgz keepalived/
# scp keepalived.tgz 192.168.1.226:/usr/local/
226服务器：
# cd /usr/local/
# tar -zxvpf keepalived.tgz
# tar -zxvpf tfs.tgz
# cat /usr/local/keepalived/etc/keepalived/keepalived.conf
global_defs {
router_id tfs_ns
}
vrrp_script chk_nameserver {
script "killall -0 nameserver"
interval 2
weight 2
}
vrrp_instance VI_1 {
state MASTER
interface eth0
virtual_router_id 23
priority 100
advert_int 1
authentication {
auth_type PASS
auth_pass password_123
}
virtual_ipaddress {
192.168.1.229/24 brd 192.168.255.255 dev eth0 label eth0:0
}
track_interface {
eth1
}
track_script {
chk_nameserver
}
}三：修改data server 配置
# grep -2 ‘vip‘ /usr/local/tfs/conf/ds.conf //两台data server做同样的修改操作 [dataserver] #nameserver ip addr(vip) ip_addr = 192.168.1.229
四：启动服务
# /usr/local/tfs/scripts/tfs start_ds 1-3 //227,228服务器上分别执行 # /usr/local/tfs/scripts/tfs start_ns //225,226服务器上分别执行 # /usr/local/keepalived/sbin/keepalived -f /usr/local/keepalived/etc/keepalived/keepalived.conf //225,226服务器上分别执行 225服务器将成为主节点，并注册vip # ip a
# tail -f /var/log/messages
226服务器将成为备份节点 # tail -f /var/log/messages
五：读写测试
# /usr/local/tfs/bin/ssm -s 192.168.1.229:8108 -i show machine
# /usr/local/tfs/bin/tfstool -s 192.168.1.229:8108 TFS> put /var/log/messages
Nginx配置文件要稍作修改，指向VIP # grep ‘8108‘ /usr/local/nginx/conf/nginx.conf server 192.168.1.229:8108; # service nginx restart http://192.168.1.12:7500/v1/tfs/T1bRxTByJT1RCvBVdK
六：failover测试
225上手动关闭name server服务 # sh /usr/local/tfs/scripts/tfs stop_ns nameserver exit SUCCESSFULLY # tail -f /var/log/messages
观察226服务器的日志，自动提升为master # tail -f /var/log/messages
重新上传，读取测试 # /usr/local/tfs/bin/tfstool -s 192.168.1.229:8108 TFS> put /etc/group
http://192.168.1.12:7500/v1/tfs/T1TRxTByZT1RCvBVdK
重新启动225上的name server服务，观察225,226的日志输出 # sh /usr/local/tfs/scripts/tfs start_ns # tail -f /var/log/messages
# tail -f /var/log/messages
本文出自 “斩月” 博客，谢绝转载！
原文地址：http://ylw6006.blog.51cto.com/470441/1559271
