Database on Café Com Cloud

A how-to from AWS

Wed, 14 Nov 2018 17:01:29 -0200

I know you heard about the Amazon outage in a specific store when migration from Oracle to Aurora. But what you didnt heard is: they already migrate 88% of their Oracle park to Aurora, and the only real issue was what came to public (of course).

So, if you wanna to check out this how to, how to use AWS DMS service, grab a coffee and follow the link below:

https://aws.amazon.com/blogs/database/best-practices-for-migrating-an-oracle-database-to-amazon-rds-postgresql-or-amazon-aurora-postgresql-migration-process-and-infrastructure-considerations/

See you Around,

18c install on premises - first impressions

Mon, 10 Sep 2018 18:35:18 -0300

Hey all,

Somethings have change on 18c. Today I will explain my first impressions of this pretty new product on premises and some comparisons by older versions (PS: I classified each topic before start talk about… I hope to help you on your next upgrade)

Download: If you use download from Oracle OTN, you notice that there is only one zip for Oracle database and one zip file for grid install.
Install: now you have to unzip all content into Oracle_HOME<- they called this new way as “Simplified Image-based Oracle Database Installation”
Install: runInstaller was replaced by gridSetup.sh (in case of Grid installation) and for database Installer you still need to execute runInstall.sh
Install: you will need at least a machine with 8 GB RAM to install grid / database.
Install: for database installer, now you have both options of install (as 10g Installer) - you can choose install enterprise or Standard Edition 2.
Install: now you can choose on database install TFA during execution of root.sh.

See you. :-)

Backup in the Cloud era - what is changing?

Wed, 20 Dec 2017 13:07:05 -0200

Hello there!

Hope this post find you well. These days one of my customers ask me to advise him on backup/restore procedures and solutions on a new environment running on AWS.

Production databases will run on EC2 instances, on Oracle with BYOL (bring your own license). Customer is thinking about EBS (Elastic Block Storage) solutions or S3 (Simple Storage Service).

Snapshots in Oracle world usually are used in conjunction with begin/end backup operations. If you take 1 snapshot per day only, and you are not willing to lost 24hs of commited data you need to have a second backup strategy. There is a product offered by N2WS which is quite amazing in orchestrating, scheduling and controlling the snapshots - you can configure to take a picture every 5 min - which is usually lower than a business RPO. But how about a logical corruption, a wrong delete done on last weekend. How to restore this? You must retain snaps for a week, maybe a month, maybe a year, restore this snap in another EC2 instance and manually restore data. Seems costly, right?

The ideal world way - from the documentation - is to deploy OSBCS Oracle Secure Backup Cloud Service. This (paid) option give you the ability to use S3 as tape, so minimal adjustments need to be done in order to migrate your backup strategy to the cloud. You need to setup and install OSBCS on each EC2 instance, adjust your channels to use OSBCS and you are ready to go! RMAN retention, catalog, everything go smoothly if you choose to pay for this option. How OSBCS is charged? Per Channel.

So, if you have 100 databases, you can buy for example 10 channels, and make 1 backup at time with 10 channels, or 10 backups at time with 1 channel, or 2 backups at time with 5 channels, or 5 backups at time with 2 channels, or - you got the idea.

Comments are welcome =]

See you around,

Hang.

Why you should, and should not, worry about the new era of inteligent databases

Sun, 24 Sep 2017 16:33:23 -0300

These days I feel like an terminator movie, with Ciberdyne working really hard to make Skynet a cient being - to become inteligent.

We all know that some activities are a best fit to automation, reason why we spend hours learning phyton or perl. If you gain your daily bread by executing scripts, loading data into the database, creating users, increasing tablespace sizes, making basic health checks - YES - you need to step up and advance in your Oracle DBA game. The machines are getting smarter on a dailly bases, you should grow smarter as well.

So, in my opinion, what activities will not be feasible to the machines in the short, perhaps medium run? Performance analysis, deep troubleshooting, installation cenarios that escape from the next-next-finish rule, Database architecture, Data migration and consolidation…things that usually people have to think and design more (using the 80/20 plan/execute rule) will not be played by the machines in the short run.

If your performance analysis in done by the dbms_sqltune - you need to worry too :)

As always feedbacks are always welcome.

See you around,

Hang.

The infamous jdbc closed connection

Mon, 11 Sep 2017 12:19:34 -0300

Sometimes you, as an DBA, is blamed for everything. The database is slow, unavailable, unpatched, and the list goes on and on.

Sometimes, in rare situations, you can prove then wrong :D

Last week we have been called to analyze a intermitent application issue. The app team blame the database, showing the “java.sql.SQLException: Closed Connection” error on app logs. Everything at database level was checked out, and rechecked again, until they call us. Long debugging hours and still no results, we tried a different approach. What about sniffing the eth card at app server..checking the communication flow between APP and Database.

And so we did it:

tcpdump -i eth0 tcp port 1521 -A -s1500 | awk ‘$1 ~ “ORA-” {i=1;split($1,t,“ORA-”);while (i <= NF) {if (i == 1) {printf("%s",“ORA-“t[2])}else {printf("%s “,$i)};i++}printf(”\n”)}’

This give us the nice output:

bla@app_blaserver:~ # tcpdump -i eth0 tcp port 1521 -A -s1500 | awk ‘$1 ~ “ORA-” {i=1;split($1,t,“ORA-”);while (i <= NF) {if (i == 1) {printf("%s”,“ORA-“t[2])}else {printf("%s “,$i)};i++}printf(”\n”)}’ tcpdump: verbose output suppressed, use -v or -vv for full protocol decode listening on eth0, link-type EN10MB (Ethernet), capture size 1500 bytes ORA-01403:no data found ORA-00913:too many values ORA-01403:no data found ORA-01403:no data found ORA-01403:no data found ORA-01403:no data found ORA-01403:no data found ORA-01403:no data found ORA-01403:no data found ORA-01403:no data found ORA-01438:value larger than specified precision allowed for this column ORA-06512:at line 2 ORA-00937:not a single-group group function ORA-01403:no data found ORA-01403:no data found ORA-00937:not a single-group group function ORA-01403:no data found ORA-00937:not a single-group group function ORA-01403:no data found ORA-00937:not a single-group group function ORA-01403:no data found

The ORA-01403 is expected after the fetch of each cursor being processed - no bad news here.

Hummmm…and when the ORA-01438/ORA-06512/ORA-00937 are raised what happens to the connection? You got it right?

After checking what was causing the errors, the intermittent issue stops, everyone was happy - incluing the DBA team :D

(You need to adapt the script to fit the listener port and eth card in your box, okay?)

As always feedbacks are very welcome.

See you around,

Hang.

12cR1 + Suse 12 SP2 = take a look on this post

Sat, 10 Jun 2017 22:19:15 -0300

Today I face a weird issue during a HAS 12.1.0.2 in Suse 12 SP2 for a customer.

Everything went smooth during the runInstaller, but roothas.pl failed with the follow error:

SERVER_BLA:/u01/grid/12.1.0.2 # /u01/grid/12.1.0.2/perl/bin/perl -I/u01/grid/12.1.0.2/perl/lib -I/u01/grid/12.1.0.2/crs/install /u01/grid/12.1.0.2/crs/install/roothas.pl Using configuration parameter file: /u01/grid/12.1.0.2/crs/install/crsconfig_params LOCAL ADD MODE Creating OCR keys for user ‘ora12c’, privgrp ‘dba’.. Operation successful. LOCAL ONLY MODE Successfully accumulated necessary OCR keys. Creating OCR keys for user ‘root’, privgrp ‘root’.. Operation successful. CRS-4664: Node SERVER_BLA successfully pinned. 2017/06/10 00:46:42 CLSRSC-330: Adding Clusterware entries to file ‘oracle-ohasd.service’

PRCR-1006 : Failed to add resource ora.ons for ons PRCR-1115 : Failed to find entities of type resource type that match filters (TYPE_NAME ends .type) and contain attributes CRS-0184 : Cannot communicate with the CRS daemon. 2017/06/10 00:47:04 CLSRSC-180: An error occurred while executing the command ‘srvctl add ons’ (error code 0)

2017/06/10 00:47:55 CLSRSC-115: Start of resource ‘ora.evmd’ failed

2017/06/10 00:47:55 CLSRSC-202: Failed to start EVM daemon

This PRCR-1006 drive me nuts. Turns out, after long hours digging the web + metalink + caffeine, that you need to:

vi /etc/ld.so.conf and add /lib64/noelision on top of it.

The file should be like this after vi:

SERVER_BLA:/u01/grid/12.1.0.2/crs/install# cat /etc/ld.so.conf /lib64/noelision /usr/local/lib64 /usr/local/lib include /etc/ld.so.conf.d/*.conf

And then, add the symlink on GI_HOME:

SERVER_BLA:/u01/grid/12.1.0.2/lib# lns -s /lib64/noelision/libpthread-2.22.so libpthread.so.0

And only after this run root.sh/roothas.pl. If you are reading this post problably you already executed roothas.pl ( =[ sadly). So you need to deconfigure first, execute the steps above, and then execute roothas.pl again.

After this, happiness is back =]

***Update***

When patching GI home, remove the symlink created, apply the patch, and make the symlink again.

Feedbacks are always welcome! See you around

Hang,

db_multiblock_read_count - to be or not to be?

Thu, 25 May 2017 18:37:21 -0300

Sometimes you wanna to use things with default options…and sometimes the default is not good enough.

We all know that db_file_multiblock_read_count is bumped by default on startup, usually you get 128 blocks per read. Why?

From Oracle documentation, this value is high, but CBO will not favor full tables scans.

“Even though the default value may be a large value, the optimizer will not favor large plans if you do not set this parameter. It would do so only if you explicitly set this parameter to a large value.”

Okay. But Why? From a 10053 trace within a session, you can see that the CBO will compute db_file_multiblock_read_count always as “8”…even if you see 128 in spfile.

CBO you always use 8, if you dont set db_file_multiblock_read_count explicitly. How to know if you set db_file_multiblock_read_count as 128? Use this:

SELECT nvl(value,’NULL’) AS value FROM v$spparameter WHERE name = ‘db_file_multiblock_read_count’;

If this return *NULL*, you have not set db_file_multiblock_read_count, Oracle is showing you 128, but CBO uses 8. Sad no?

You all read the documentation and know the effects of bumping db_file_multiblock_read_count high. Full table scans will happen more often and as a result, in a OLTP system, usually you will get a call from your boss (hehe)

But there is a way to set correctly the db_file_multiblock_read_count…by testing…for example, take a look on the block below. Its silly, I know, but it can give you a magic value for this parameter.

set serveroutput on DECLARE l_count PLS_INTEGER; l_time PLS_INTEGER; l_starting_time PLS_INTEGER; l_ending_time PLS_INTEGER; BEGIN dbms_output.put_line(‘dbfmbrc seconds’); FOR l_dbfmbrc IN 1..64 LOOP EXECUTE IMMEDIATE ‘ALTER SESSION SET db_file_multiblock_read_count=’||l_dbfmbrc; EXECUTE IMMEDIATE ‘ALTER system flush buffer_cache’; EXECUTE IMMEDIATE ‘ALTER session disable parallel dml’; l_starting_time := dbms_utility.get_time(); SELECT /+ full(t) / count(*) INTO l_count FROM big_table t; l_ending_time := dbms_utility.get_time(); l_time := round((l_ending_time-l_starting_time)/100); dbms_output.put_line(l_dbfmbrc||’ ‘||l_time); END LOOP; END; /

(Thanks to Chris Antognini, you’re great)

With this block (be carefull, it will take 60-90 minutes depending of the big table’s size, as a advice, use a 5GB table), you will have values, plot a graph and see the behavior. Dont choose a high value for OLTP.

See you around!

12c Upgrade Exam - 1z0-060

Wed, 12 Apr 2017 13:25:10 -0300

Today we will describe some impressions about 1z0-060 exam.

So, for get success in exam, you need:

Study hard on 12c architecture (CDB / PDB): how to start/ stop, create, drop, secure, etc.
Do not believe on dumps answers.

We recommend to do new features course, or have some experience on 12c first - some questions have direct relation with daily DBA tasks.

First impression on 12cR2 database install

Thu, 16 Mar 2017 16:00:00 -0300

Today I’ll describe my first impressions on install the new 12cR2 (available for premises on march 2017) for linux and Solaris.

Inicially I felt a little bit strange because there are only one file for database and only one for grid install (on other versions since 11gR2, there are 2 files for database and one or more for grid - depends of version that you install).

So I downloaded and install it on OEL6.8 for study purposes.

The install process is almost similar than 12cR1 process, except you need more free space to install it (at least 8 GB).

Other point, when you unzip the single file, keep careful in uncompress on a folder (for organization, because all files will be uncompress in same directory that you are).

Then I will keep study and publish some more new features! :-)

To be or not to be (lazy)

Mon, 06 Mar 2017 20:49:44 -0300

Sometimes you are an Oracle DBA just tired of perform setup ( I like to install everything from scratch…since oracle 7.3 it was always an adventure for me :) )

If you only want to download and have fun on some cool new features, a pre-installed ova is the gold medal for you..

Check out last this Jeff Smith (SQL Developer and lots of fun on his blog) here.

Its a OVA with 12.2, APEX, SQL Developer and other cool stuff :)

See you around!

Database on Café Com Cloud

A how-to from AWS

18c install on premises - first impressions

Backup in the Cloud era - what is changing?

Why you should, and should not, worry about the new era of inteligent databases

The infamous jdbc closed connection

12cR1 + Suse 12 SP2 = take a look on this post

2017/06/10 00:47:55 CLSRSC-115: Start of resource ‘ora.evmd’ failed

2017/06/10 00:47:55 CLSRSC-202: Failed to start EVM daemon

SERVER_BLA:/u01/grid/12.1.0.2/crs/install# cat /etc/ld.so.conf /lib64/noelision /usr/local/lib64 /usr/local/lib include /etc/ld.so.conf.d/*.conf

SERVER_BLA:/u01/grid/12.1.0.2/lib# lns -s /lib64/noelision/libpthread-2.22.so libpthread.so.0

db_multiblock_read_count - to be or not to be?

SELECT nvl(value,’*NULL*’) AS value FROM v$spparameter WHERE name = ‘db_file_multiblock_read_count’;

12c Upgrade Exam - 1z0-060

First impression on 12cR2 database install

To be or not to be (lazy)

SELECT nvl(value,’NULL’) AS value FROM v$spparameter WHERE name = ‘db_file_multiblock_read_count’;