Chinaunix

标题: [原创]NBU 的tune [打印本页]

作者: alex_linux 时间: 2006-10-23 14:36
标题: [原创]NBU 的tune
这几天user的nbu很慢,想作些tune,我也做了些准备和试验,和大家分享一下.\r\nNbu tune 从大的说很复杂\r\n\r\n一硬件\r\n\r\n主要是架构,机器的选型(这个先不说,环境要求不同,差距很大)\r\n\r\n二软件\r\n\r\n1 nbu软件\r\n\r\n主要从几个方面分别介绍下:\r\n\r\nTCP/IP Network Buffer Size.\r\n\r\n/usr/openv/netbackup/NET_BUFFER_SZ is a file containing a number indicating the TCP/IP socket buffer size that should be used for data transfers between the NetBackup media server and its clients. If the file does not exist, the default value used is 32032 bytes.\r\n\r\nThe default value can be changed by creating the file /usr/openv/netbackup/NET_BUFFER_SZ on the NetBackup media server and client systems.\r\n\r\nThe /usr/openv/netbackup/NET_BUFFER_SZ file is a text file containing a single line specifying the value of the Network Buffer Size in bytes in decimal format.\r\n\r\nFor example, a Network Buffer Size of 64KB could be configured and checked like this:\r\n\r\n# echo \"65536\" > /usr/openv/netbackup/NET_BUFFER_SZ\r\n# cat /usr/openv/netbackup/NET_BUFFER_SZ\r\n65536\r\n# \r\n\r\nSpecifying a larger value in the NET_BUFFER_SZ file may improve the performance of backups and restores. It is recommended that if NET_BUFFER_SZ is used, the same value should be set on all the NetBackup media servers and clients.\r\n\r\n\r\n\r\nData Buffer Size and Number of Data Buffers.\r\n\r\n\r\n\r\nThe NetBackup media server uses shared memory to buffer data between the network and the tape drive (or between the disk and the tape drive if the NetBackup media server and client are the same system). By default, NetBackup uses a default value of 8 x 32KB shared memory buffers for non-multiplexed backups and 4 x 64KB for a multiplexed backup. \r\n\r\nThese buffers can be configured by creating the files /usr/openv/netbackup/db/config/SIZE_DATA_BUFFERS and /usr/openv/netbackup/db/config/NUMBER_DATA_BUFFERS on the NetBackup media server. These parameters are meant exclusively for media servers, and should not be used on a pure master server or client.\r\n\r\nThe SIZE_DATA_BUFFERS file should contain a single line specifying the data buffer size in bytes in decimal format.\r\nThe NUMBER_DATA_BUFFERS file should contain a single line specifying the number of data buffers in decimal format.\r\n\r\nThe value in the SIZE_DATA_BUFFERS file has to be a multiple of 1024B.\r\nIt is common to increase SIZE_DATA_BUFFERS to 256KB and NUMBER_DATA_BUFFERS to 16.\r\n\r\n\r\nIMPORTANT: Because the data buffer size equals the tape I/O size, the value specified in SIZE_DATA_BUFFERS must not exceed the maximum tape I/O size supported by the tape drive or operating system. This is usually 256 KB or 128 KB.\r\n\r\n\r\nIn order to configure NetBackup to use 16 x 256 KB data buffers, specify 262144 (256 x 1024) in SIZE_DATA_BUFFERS and 16 in NUMBER_DATA_BUFFERS:\r\n\r\n# echo \"262144\" > /usr/openv/netbackup/db/config/SIZE_DATA_BUFFERS\r\n# echo \"16\" > /usr/openv/netbackup/db/config/NUMBER_DATA_BUFFERS\r\n# cat /usr/openv/netbackup/db/config/SIZE_DATA_BUFFERS\r\n262144\r\n# cat /usr/openv/netbackup/db/config/NUMBER_DATA_BUFFERS\r\n16\r\n# \r\n\r\nCare should be taken when changing these values. For example, increasing the tape buffer size can cause some backups to run slower and there have been cases where there have been restore problems.\r\n\r\n\r\nIMPORTANT: After making any changes, it is vitally important to verify that the following tests can be completed successfully:\r\n\r\n1. Run a backup\r\n2. Restore data from the new backup created in step 1\r\n3. Restore data from a backup created prior to the changes to SIZE_DATA_BUFFERS and NUMBER_DATA_BUFFERS\r\n\r\n\r\nThe number and size of the data buffers used by the NetBackup media server can be checked by looking in the bptm log file: \r\n\r\n12:02:55 [28551] <2> io_init: using 32768 data buffer size\r\n12:02:55 [28551] <2> io_init: CINDEX 0, sched bytes for monitoring = 2000\r\n12:02:55 [28551] <2> io_init: using 8 data buffers\r\n\r\nor\r\n\r\n15:26:01 [21544] <2> mpx_setup_restore_shm: using 12 data buffers, buffer size is 65536\r\n\r\n\r\nIncreasing the size and number of the data buffers will use up more shared memory, which is a limited system resource. The total amount of shared memory used is: (buffer_size * num_buffers) * drives * MPX. \r\n\r\nSo, for two tape drives, each configured with an MPX of 4, and with 16 data buffers of 256 KB, the total shared memory usage would be:\r\n\r\n(16 * 256k) * 2 * 4 = 32 MB (32768 KB)\r\n\r\nTake into consideration the total system resources and the entire network.\r\nThe Maximum Transmission Unit (MTU) for the LAN may also have to be changed.\r\n\r\n\r\nNOTE: For Digital Linear Tape (DLT) drives, the number that seems to give the best performance for SIZE_DATA_BUFFERS is 65536. Further performance improvements can be achieved by increasing the value in the NUMBER_DATA_BUFFERS file.\r\n\r\n\r\nNUMBER_DATA_BUFFERS_DISK\r\n\r\n和 NUMBER_DATA_BUFFERS\r\n的用法相似,但,一般瓶颈不会再这里,所以不在重复\r\n\r\n\r\n\r\n2 Os kernel tune(unix)\r\n\r\nOs 这里主要以solaris为例(比较熟)\r\nTune 也主要是 share memory\r\n下面是一些参数和推荐tune\r\n\r\n■ Message queues\r\n\r\nset msgsys:msginfo_msgmax = maximum message size\r\nset msgsys:msginfo_msgmnb = maximum length of a message queue in\r\nbytes. The length of the message queue is the sum of the lengths of all the\r\nmessages in the queue.\r\nset msgsys:msginfo_msgmni = number of message queue identifiers\r\nset msgsys:msginfo_msgtql = maximum number of outstanding messages\r\nsystem-wide that are waiting to be read across all message queues.\r\n\r\n■ Semaphores\r\n\r\nset semsys:seminfo_semmap = number of entries in semaphore map\r\nset semsys:seminfo_semmni = maximum number of semaphore identifiers\r\nsystem-wide\r\nset semsys:seminfo_semmns = number of semaphores system-wide\r\nset semsys:seminfo_semmnu = maximum number of undo structures in\r\nsystem\r\nset semsys:seminfo_semmsl = maximum number of semaphores per id\r\n\r\nset semsys:seminfo_semopm = maximum number of operations per semop\r\ncall\r\nset semsys:seminfo_semume = maximum number of undo entries per\r\nprocess\r\n\r\n■ Shared memory\r\n\r\nset shmsys:shminfo_shmmin = minimum shared memory segment size\r\nset shmsys:shminfo_shmmax = maximum shared memory segment size\r\nset shmsys:shminfo_shmseg = maximum number of shared memory\r\nsegments that can be attached to a given process at one time\r\nset shmsys:shminfo_shmmni = maximum number of shared memory\r\n\r\n\r\n\r\n*BEGIN NetBackup with the following recommended minimum settings in a\r\nSolaris /etc/system file\r\n*Message queues\r\nset msgsys:msginfo_msgmap=512\r\nset msgsys:msginfo_msgmax=8192\r\nset msgsys:msginfo_msgmnb=65536\r\nset msgsys:msginfo_msgmni=256\r\nset msgsys:msginfo_msgssz=16\r\nset msgsys:msginfo_msgtql=512\r\nset msgsys:msginfo_msgseg=8192\r\n*Semaphores\r\nset semsys:seminfo_semmap=64\r\nset semsys:seminfo_semmni=1024\r\nset semsys:seminfo_semmns=1024\r\nset semsys:seminfo_semmnu=1024\r\nset semsys:seminfo_semmsl=300\r\nset semsys:seminfo_semopm=32\r\nset semsys:seminfo_semume=64\r\n*Shared memory\r\nset shmsys:shminfo_shmmax=16777216\r\nset shmsys:shminfo_shmmin=1\r\nset shmsys:shminfo_shmmni=220\r\nset shmsys:shminfo_shmseg=100\r\n*END NetBackup recommended minimum settings\r\n\r\n以上只一部 nbu的tune 还是很多,很复杂的.

作者: alex_linux 时间: 2006-10-23 19:43
嘿嘿,,很懒的写东西,,,,,,,,,,

作者: liwenguo 时间: 2006-10-24 21:28
提示: 作者被禁止或删除内容自动屏蔽

作者: alex_linux 时间: 2006-10-25 09:37

原帖由 liwenguo 于 2006-10-24 21:28 发表\r\n搂主是veritas的手册吗？能告诉是哪一本？

\r\n不是..谢谢.

作者: maping 时间: 2006-10-28 18:38
请楼主继续，写更好的帖子。

作者: alex_linux 时间: 2006-10-30 10:29

原帖由 maping 于 2006-10-28 18:38 发表\r\n请楼主继续，写更好的帖子。

\r\n其实不想写

\r\n老大，逼着写report啊。。。\r\n\r\n应该会有后续吧。。。不确定了

作者: filwy 时间: 2006-10-30 11:10
请问问，nbu能把备份的数据还原的吗？\r\n有没有还原的功能的呢。还是要自己写上脚本去还原的？？

作者: alex_linux 时间: 2006-10-30 12:12

原帖由 filwy 于 2006-10-30 11:10 发表\r\n请问问，nbu能把备份的数据还原的吗？\r\n有没有还原的功能的呢。还是要自己写上脚本去还原的？？

\r\n

作者: maping 时间: 2006-10-30 14:09
备份的数据当然可以还原.\r\n也可以用tar把磁带中的数据读出来.

作者: starmoon2003 时间: 2006-10-30 20:45
标题: 上面的TUNE 我已经在项目中用过
速度可以提高3倍以上。

作者: maping 时间: 2006-10-30 22:17
不要只追求备份的速度而放弃了系统的性能。\r\n建议速度提高后，备份完成后，做一个恢复测试。

作者: 笑看风云_淡 时间: 2006-10-31 09:50
还是看原厂的文档好.本来veritas的一般文档都share的\r\nhttp://ftp.support.veritas.com/p ... _Enterprise_Server/

作者: heroesray 时间: 2007-01-16 17:13
标题: 回复 1楼 alex_linux 的帖子
SIZE_DATA_BUFFERS等好像只适用于nbu 5.0吧,6.0没有/usr/openv/netbackup/db/config/这个目录,去年9月装了个hp-ux的,想优化结果没这个目录,打支持电话也没什么结果,只好不优化了.

作者: alex_linux 时间: 2007-01-16 19:09

原帖由 heroesray 于 2007-1-16 17:13 发表\r\nSIZE_DATA_BUFFERS等好像只适用于nbu 5.0吧,6.0没有/usr/openv/netbackup/db/config/这个目录,去年9月装了个hp-ux的,想优化结果没这个目录,打支持电话也没什么结果,只好不优化了.

\r\n我的6.0\r\n有 ,莫非我装错了

作者: henrypan 时间: 2007-04-29 20:45
标题: Cool
Thanks to Lazy boy:>)

作者: qqadmini 时间: 2007-11-13 16:46
顶。。眼熟\n\n[ 本帖最后由 qqadmini 于 2007-11-13 16:49 编辑 ]

欢迎光临 Chinaunix (http://bbs.chinaunix.net/)