Introduction
Recently our new association DBASK small operation and maintenance procedures of the United States, the number of high-end storage public knowledge, a coffee Sen Hutchison, operation and maintenance coffee bar and other data fields, welcome to read to share.
Q & A Highlights
Next, we share this issue sorted out the problems and diagnostic summary, for reference study, a detailed analysis of the diagnostic process can jump to the small program to view by title link.
One problem, temp table space for the full situation, how to deal with?
great undo temp table space, my idea is to increase the table space size or create a new one temp1, then default to develop new temp1, but I feel this will be full. Normal idea how to deal with this undo temp table space is a big problem?
Conclusion The diagnosis: the most fundamental way is to optimize occupancy and undo many temp sql, such as sorting and more, chief executive of time sql and so on. Combined with their environment and set a reasonable temp undo space, usually to the core business users separate temp table space, so as not to affect other misuse of the core business. In addition these two table space is not released, is running with the usage of business will be more and more, and finally close to 100%, generally if not reported ora-01555 error and expansion temp table space can not be ignored table space usage.
Question two, extended cluster when a site failure, the processing of the process instance in the survival sites
survival process instance handle this failure site, as well as general estimate how long it takes to restore service capabilities.
Diagnostic conclusion: an example of two typical faults, 1, both sides of the bearer service database, fails either side, by the other side to carry all without the need for switching the primary and secondary sites. User session failed site previously hosted taken over in a few minutes by the other surviving node, there will be some Caton on experience, the same as ordinary RAC node failure, an instance fails, the cluster reconfiguration, application TAF. 2, if the heartbeat between sites is interrupted, the third disc by the arbitration decision which site expel the site, which site alive. Expelled former site user experience will be hosted on a few minutes of Caton.
Question three, Oracle 12c OEM upgrade to 13c
upgrade the local server agent process.
Conclusion The diagnosis: This depends on the specific version, even if all you need to upgrade from 12.1.0.1 to 12.1.0.5 to 12.1.0.3 and then upgrade to 12.1.0.5,12 to support a minimum of 13 direct upgrade version 12.1.0.5, other needs first upgrade to an intermediate version. After completion of GC upgrade, upgrade agent directly to the GC management page.
Question four, Oracle Database first visit stuck
place backstage connect oracle database system page, first access the system when the page frame can come out, but the need to tune the database to display data has been turning in circles, not an error, refresh the page again when the data came out, it was to ask what parameters affect the database do?
Conclusion The diagnosis: This still have to start from the application, debug open look at what went wrong. For databases, there is just the first execution of SQL need to parse, read disk data cached in memory, the second performance will take data directly from memory, the speed will be faster. Also this phenomenon is likely to be connected to a long connection to the database is not carried out due to prolonged data exchange, gateway or firewall is turned off, so the first visit of TNS will wait for a timeout. It allows applications to increase output log and see if the problem takes place in the first run after a SQL database connection. If this is the problem, solve the idea is to modify the configuration of a gateway or firewall, to avoid interruption of idle connections. Or dead in the database configuration connection check DCD.
Five questions, uninstalled DB, how to completely uninstall GI
Uninstall grid, run the second root script error, presumably because asm instance is not close to a must have to keep open. Asm hand close all instances run with sysasm or error, perform ./crsctl stat res -t found .dg .proxy_advm .asm and other resources, although it ora.ASMNET1LSNR_ASM.lsnr becomes offline or online, crsctl stop crs can not perform how to solve ?
Conclusion The diagnosis: attempt to forcibly stop: crsctl stop crs -f, then deconfig -force and deinstall, detailed steps, please see the details.
Question six, 9.2.0.8 aix upgrade need to check what the 11.2.0.4 aix upgrade
refer to "manually upgrade to Oracle Database 11gR2 (11.2) complete checklist (document ID 1674333.1)" upgrade, which timezone does not upgrade. In addition to checking the upgrade log, background log, the user object after the upgrade.
Conclusion The diagnosis: optimizer from RBO to the CBO, changed a lot, to ensure that the application before the production database upgrade has been done fully SPA test.
Question seven, win install 10g-12546 TNS error Ora: permission denied
Windows 2012 r2 64 64 10.2.0.4 database system is installed, there will always appear this error, either dbca time, or is completely build a database, a server restart after, sqlplus also can not login error. Try a little patch 10.2.0.5 installed, then no problem in the event, repeated the test several times, this is a bug it?
诊断结论：这个是因为Oracle 10.2.0.4发行的时候，Windows 7都还没有出，而从Windows 7开始（甚至是更早的Windwos Vista开始），Windows的权限体系出现了非常大的变化，而早期的Oracle版本并不能适应这种新环境，所以会出各种各样的问题。用新版本就好了，如你说的Oracle 10.2.0.5，该小版本正好解决这个环境问题，所以就没问题了。
问题八、oracle数据库平台兼容性列表
Oracle各个版本操作系统兼容性列表
诊断结论：可以参考https://docs.oracle.com/en/database/oracle/oracle-database/index.html，每个版本都有不同的支持列表，在官方文档上可以查到。
问题九、dg库控制文件扩展报错
dg库的control1扩展报错，用的是裸设备，2和3还是原来大小，主库的没有报错，并且大小没有变，求解释。
专家解答：从ALERT报错可以看出，控制文件从8050个块扩展到13168个块时报错，而裸设备最大只支持8050个块，无法扩展，可以尝试将参数CONTROL_FILE_RECORD_KEEP_TIME改小，避免控制文件报错。
问题十、sql调条件中where1=1，基于不同的优化器cbo会走索引，而rbo则不会
应用中经常能够看到where1=1的条件。请问，为什么在不同的优化器下，rbo不会走索引而是全扫？
诊断结论：与where 1=1没有关系，这个SQL相当于没有where条件。RBO是定义死的一套规则，只有在MAX和MIN索引列时才会走索引全扫，也不会去看统计信息。
问题十一、OGG 11.2无法使用format release 12.3
高版本OGG传递文件到11.2，传递失败，试过加参数format release，不行，高版本OGG传递文件到12.1，应用失败，试过加参数format release，不行。求一份OGG版本兼容。
诊断结论：建议源端、目标端最好用同一版本。目前从官方文档来看，10、11、12都可以通过FORMAT RELEASE参数来兼容，除了CDB必须要求12.3以上外。
想了解更多知识点吗？
那就登录“墨天轮”了解学习吧~