玉面飞龙的BLOG

带宝宝出来见见世面，免得呆呆傻傻。
Read more…

By sandeep makol

Yong Huang shared a link: “Took me a while to find the orig author of “Understanding the SCN”. It’s Geoff Ingram, in his 2002 book “High-performance Oracle”. You can see a paragraph on books.google.com. Shame on the plagiarist at http://www.dbapool.com/articles/1029200701.html.”

Read more…

前一阵子宝宝生病了，发烧，最高39度，最低38度，虽然经过物理降温，但体温反复好几次，吐奶，无精神。去医院急诊，一个有经验的大夫问道，这是孩子第一次发烧吧，如果是的话，就不是大问题，基本上每个宝宝，6个月大之后，由于母乳喂养，出生时候携带的抵抗力会慢慢消失，6个月到3岁大的时候容易感染病毒，导致发烧。

该发烧会反复，吃什么药都不管用，发烧3天后，温度会自动降低到正常，并且身上会起一些小的红色的疹子；出了疹子之后发烧就会好，疹子也会在1到天后自动消失。这是宝宝增强自身抵抗力的过程。但如果宝宝连续发烧超过3天，则要留意医院就诊最好。

特此提醒广大新生儿家长，宝宝第一次发烧的时候要淡定。

上周给同事们做过一个redis的功能介绍。PPT这里提供下载redis-io-copy。该ppt里收录了很多与redis相关的一些资料和链接。

Redis是一款开源nosql key-value数据库产品；，使用redis的开发人员很多，它的value可以是几种数据结如数组，链表，因此也常被使用者称为数据结构key-value数据库。其作者和社区都很活跃，得益于其短小精悍，容易上手，发展迅猛，而且国内也有不少标杆用户，如weibo和taobao等。

btw:NoSQLFan是个好网站，至于你信不信，我反正信了

Oracle 11g introduced V$SESSION.SQL_EXEC_START column. It’s “Time when the execution of the SQL currently executed by this session started; NULL if SQL_ID is NULL”.

Plus LAST_CALL_ET,it’s easier to identify session status and detect problem SQL.

SQL>   select sid,serial#,LAST_CALL_ET,sql_hash_value,status,(SYSDATE-SQL_EXEC_START)*86400  from v$session where status=’ACTIVE’ and SQL_EXEC_START is not null order by (SYSDATE-SQL_EXEC_START)*86400 desc;

SID    SERIAL# LAST_CALL_ET SQL_HASH_VALUE STATUS   RUN_SECONDS
———- ———- ———— ————– ——– ———–
15022      43583            1      404603867 ACTIVE            31
15404      28683           27     3096975108 ACTIVE            26
28569       3837           26     3096975108 ACTIVE            26
28834      28447           26     3096975108 ACTIVE            26
28440      57479           26     3096975108 ACTIVE            26
28487      43835           26     3096975108 ACTIVE            26
28477       1469           26     3096975108 ACTIVE            26
28448      44175           26     3096975108 ACTIVE            26
17793      49565            9     3770652317 ACTIVE             9
28602      12903            7     1278135340 ACTIVE             7
25851         25              5     4149984594 ACTIVE             5

2 way to re-sync one out-of-sync table in Oracle logical standby.

Way 1:

sql> alter database stop logical standby apply ;
sql> exec dbms_logstdby.instantiate_table(‘SCOTT’,’EMP’,’DBLINK’);
sql> alter database start logical standby apply immediate;

The above command also does expdp/impdp using the network_link which is the db link.But it will drop the existing out-of-sync table first, then creates and loads it with the data the source using the db link.

Way 2:

manually expdp & impdp out-of-sync tables from master to logical standby.

sql> alter database stop logical standby apply;
sql> exec dbms_logstdby.skip (stmnt =>'DML',schema_name =>'SCOTT',table_name =>'EMP');
sql> drop table scott.emp
sql> create table scott.emp (ename varchar(2),........);
sql>SELECT APPLIED_SCN, LATEST_SCN, MINING_SCN, RESTART_SCN FROM V$LOGSTDBY_PROGRESS;

APPLIED_SCN :-All the transactions with COMMIT SCN lower than or equal to this SCN have been applied.

then on Primary, do flashback expdp with SCN=APPLIED_SCN.

expdp system/xxxx tables=scott.emp dumpfile=expemp.dmp flashback_scn=<APPLIED_SCN>
impdp system/xxxx dumpfile=expemp.dmp tables=scott.emp table_exists_actions=append

you can expdp rows only, then create index manually with large sort_area_size and parallel.

Everything is done so far. Start apply.

sql> exec dbms_logstdby.unskip(stmnt =>’DML’,schema_name =>’SCOTT’,table_name =>’EMP’);
sql> alter database start logical standby apply;

Some case , there’re high waits on event ‘resmgr:cpu quantum’, while there is no CPU load on the host.

Resmgr means it’s related to oracle resource manager, but if you’ve disabled it in past, why does it come from?

You should “ALTER SYSTEM SET RESOURCE_MANAGER_PLAN = 'FORCE:' ” to deeply disable it or use underscore parameter _resource_manager_always_on = false.
Read more…

原来卖磁盘的销售人员也卖过骆驼的。

骆驼买３个，数据存３份，孩子养３个。

Read more…