fastdfs 上传测试，请版主帮看看。

squall1

两台tracker     192.168.8.90/91
两台storage     192.168.8.92/93
一台client      192.168.8.94

测试场景一：

两台tracker同时工作，两台storage同时工作，上传1000个100K的文件。

[2011-03-18 15:45:35] INFO - base_path=/home/fastdfs_storge, connect_timeout=30, network_timeout=60, tracker_server_count=2, anti_steal_token=0, anti_steal_secret_key length=0

tracker_query_storage_store_list_without_group:
        server 1. group_name=group1, ip_addr=192.168.8.92, port=23000
        server 2. group_name=group1, ip_addr=192.168.8.93, port=23000

group_name=group1, ip_addr=192.168.8.92, port=23000
storage_upload_by_filename
group_name=group1, remote_filename=M00/00/0F/wKgIXE2DDhyc0yafAAGQAJ2JkdA367.img
source ip address: 192.168.8.92
file timestamp=2011-03-18 15:47:40
file size=102400
file crc32=2643038672
file url: http://192.168.8.90/group1/M00/0 ... fAAGQAJ2JkdA367.img
storage_upload_slave_by_filename
group_name=group1, remote_filename=M00/00/0F/wKgIXE2DDhyc0yafAAGQAJ2JkdA367_big.img
source ip address: 192.168.8.92
file timestamp=2011-03-18 15:47:40
file size=102400
file crc32=2643038672
file url: http://192.168.8.90/group1/M00/0 ... QAJ2JkdA367_big.img

real    2m3.303s
user    0m0.584s
sys     0m25.131s
[root@client ~]#

==================================================================================================

测试场景二：

上传一半，关掉一台tracker（192.168.8.90），两台storage同时工作，上传1000个100K的文件。

关掉同时，上传文件一直较慢。上传文件的URL地址从原先的90改为91

[2011-03-18 16:47:53] ERROR - connect to 192.168.8.90:22122 fail, errno: 113, error info: No route to host
source ip address: 192.168.8.93
file timestamp=2011-03-18 16:50:24
file size=102400
file crc32=2643038672
file url: http://192.168.8.91/group1/M00/0 ... QAJ2JkdA701_big.img
This is FastDFS client test program v2.08

Copyright (C) 2008, Happy Fish / YuQing

FastDFS may be copied only under the terms of the GNU General
Public License V3, which may be found in the FastDFS source kit.
Please visit the FastDFS Home Page http://www.csource.org/
for more detail.

[2011-03-18 16:47:53] INFO - base_path=/home/fastdfs_storge, connect_timeout=30, network_timeout=60, tracker_server_count=2, anti_steal_token=0, anti_steal_secret_key length=0


[1]+  Stopped                 /bin/bash upload.sh

real    56m49.310s
user    0m0.000s
sys     0m0.015s
[root@client ~]#

花费56秒未传完。速度太慢，直接终止了。

liang3391

顶

squall1

也就是说如果坏了一台tracker，性能几乎没有了。这个如何解决？

squall1

connect_timeout=5
network_timeout=5

都设置的5秒

happy_fish100

回复 3# squall1

这是客户端API中反复对下线的tracker server进行尝试连接导致的吧？

squall1

我是用的fdfs_test  测试的。

没有采用API接口。

happy_fish100

回复 6# squall1

fdfs_test也是调用的C Client API啊！