表空间快照中direct_writes的含义 [英] Meaning of direct_writes in tablespace snapshot
问题描述
您好,
我想知道表格空间快照中direct_writes的含义
表空间上的
:
*短信(db2empfa ON)
*在INSERT中只访问1个表 - Z锁定(未记录初始模式)
*否BLOB,表中的长varchar ...
*没有实用程序运行
有人可以帮助理解我们为什么会看到大量的direct_writes吗?
版本:DB2 UDB V7 EEE FP11 / Aix 5.1 ML4
感谢您的帮助,
Jean-Marc
Hi there,
I would like to know the meaning of direct_writes in a tablespace snapshot
on a tablespace that is:
* SMS (db2empfa ON)
* Only 1 table is accessed in INSERT - Z locked (Not logged Initially mode)
* No BLOB, long varchar ...in the table
* No utility running
Can someone help understand why we see a high number of direct_writes ?
Version: DB2 UDB V7 EEE FP11 / Aix 5.1 ML4
Thanks for your help,
Jean-Marc
推荐答案
" Jean-Marc Blaise" <无**** @ nowhere.com>在留言中写道
新闻:cc ********** @ news-reader2.wanadoo.fr ...
"Jean-Marc Blaise" <no****@nowhere.com> wrote in message
news:cc**********@news-reader2.wanadoo.fr...
你好,
我想知道表空间中的direct_writes在表空间中的含义:
* SMS(db2empfa ON)
*只有1个表是在INSERT中访问 - Z锁定(未记录最初
模式)*表中没有BLOB,长varchar ...
*没有实用程序运行
有人可以帮助理解我们为什么看到大量的direct_writes?
版本:DB2 UDB V7 EEE FP11 / Aix 5.1 ML4
感谢您的帮助,
Jean-Marc
Hi there,
I would like to know the meaning of direct_writes in a tablespace snapshot
on a tablespace that is:
* SMS (db2empfa ON)
* Only 1 table is accessed in INSERT - Z locked (Not logged Initially mode) * No BLOB, long varchar ...in the table
* No utility running
Can someone help understand why we see a high number of direct_writes ?
Version: DB2 UDB V7 EEE FP11 / Aix 5.1 ML4
Thanks for your help,
Jean-Marc
这是一个大规模的插入工作吗?对于初学者来说,如果你有多个CPU,那么
会增加页面清洁工和I / O服务器的数量。你的缓冲区有多大?
池?
Is this a massive insert job? For starters, if you have multiple CPU''s,
increase number of pages cleaners and I/O servers. How large is your buffer
pool?
" Mark A" <无**** @ nowhere.com>一个écritdansle message de
新闻:Qk ***************** @ news.uswest.net ...
"Mark A" <no****@nowhere.com> a écrit dans le message de
news:Qk*****************@news.uswest.net...
Jean-Marc Blaise <无**** @ nowhere.com>在消息中写道
新闻:cc ********** @ news-reader2.wanadoo.fr ...
"Jean-Marc Blaise" <no****@nowhere.com> wrote in message
news:cc**********@news-reader2.wanadoo.fr...
你好,
我想知道表空间中的表空间
快照中direct_writes的含义:
* SMS(db2empfa ON)
*在INSERT中只访问1个表 - Z锁定(未记录初始模式)
Hi there,
I would like to know the meaning of direct_writes in a tablespace snapshot on a tablespace that is:
* SMS (db2empfa ON)
* Only 1 table is accessed in INSERT - Z locked (Not logged Initially mode)
*表中没有BLOB,长varchar ...
*没有实用程序运行
有人可以帮助理解我们为什么看到大量的direct_writes?
版本:DB2 UDB V7 EEE FP11 / Aix 5.1 ML4
感谢您的帮助,
Jean-Marc
* No BLOB, long varchar ...in the table
* No utility running
Can someone help understand why we see a high number of direct_writes ?
Version: DB2 UDB V7 EEE FP11 / Aix 5.1 ML4
Thanks for your help,
Jean-Marc
这是一个大规模的插入工作吗?对于初学者来说,如果你有多个CPU,那么
会增加页面清洁工和I / O服务器的数量。你的
Is this a massive insert job? For starters, if you have multiple CPU''s,
increase number of pages cleaners and I/O servers. How large is your
缓冲池有多大?
buffer pool?
是的,这是一个大规模的插入工作,大约有1亿行。我们每个分区有大约8
页面清理器和I / O服务器(SMS每个
分区有4个容器),还有一个ps -ef | grep -e db2clnr -e db2pfchr显示某些
进程没有消耗CPU。缓冲池大约是20000个16K页面和
chgnpgs_threshold设置为20%。
你的意思是资源短缺会导致DB2使用直接写入吗?
问候,
JM
Yes, this is a massive insert job, about 100 million rows. We have about 8
page cleaners and I/O servers per partition (SMS has 4 containers per
partition), and a ps -ef |grep -e db2clnr -e db2pfchr shows that some
processes have not consumed CPU. Buffer pool is about 20000 16K pages and
chgnpgs_threshold is set to 20%.
Do you mean that shortage of resources could lead DB2 to use direct writes ?
Regards,
JM
> >这是一个大规模的插入工作吗?对于初学者,如果你有多个CPU,
> > Is this a massive insert job? For starters, if you have multiple CPU''s,
增加页面清洁工和I / O服务器的数量。你的缓冲区有多大
increase number of pages cleaners and I/O servers. How large is your buffer
pool?
是的,这是一个大规模的插入工作,大约有1亿行。我们每个分区有大约8页清理工具和I / O服务器(SMS每个分区有4个容器),而ps -ef | grep -e db2clnr -e db2pfchr显示有些
进程没有消耗CPU。缓冲池大约是20000个16K页面,而chgnpgs_threshold被设置为20%。
你的意思是资源短缺会导致DB2使用直接写入
Yes, this is a massive insert job, about 100 million rows. We have about 8
page cleaners and I/O servers per partition (SMS has 4 containers per
partition), and a ps -ef |grep -e db2clnr -e db2pfchr shows that some
processes have not consumed CPU. Buffer pool is about 20000 16K pages and
chgnpgs_threshold is set to 20%.
Do you mean that shortage of resources could lead DB2 to use direct writes
?
问候,
JM
?
Regards,
JM
看看这篇关于插入性能的文章:
http://www-106.ibm.com/developerwork。 ..m-0403wilkins /
这篇关于表空间快照中direct_writes的含义的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!