Three SCAN IPs Resolving 8 Node RAC |
Vložit
- čas přidán 8. 07. 2024
- In this episode of daily DBA, I pick up 6 important #DBA related questions and give my answers! Do not forget to checkout BONUS QUESTION at the end of the video!
00:00 Intro
0:38 By mistake someone removed online redo logs and cleanup job has cleaned archives and we do not have any database backup scheduled at the server. What are ways to recover database again ?
03:48 What is the difference between checkpoint and SCN?
05:53 What is the difference between local and dictionary tablespace?
08:30 I want to clone database from source to Target but there is no connectivity between servers. How will clone happen with out connectivity and could you please share high level steps?
08:42 #dba Challenge!
10:45 Can you explain how 3 scan IPs resolve 8 node RAC in Oracle because we have 8 nodes in the cluster and only 3 scan IPs. How is it possible to resolve these 8 nodes?
15:58 Oracle DB has OMF. someone deleted all the control files. OS -Oracle Linux 7, DB 12c and OS already restarted. Archive log mode is not enabled. Without using the RMAN how to recover the database?
Bonus Question
20:49 What is the difference between courses and real time projects?
Website: www.dbagenesis.com/
Facebook: / dbagenesis
Instagram: / dbagenesis
Twitter: / dbagenesis - Věda a technologie
Hi Arun i do faced this archive loss issue. 1)I have modified the pfile with "_allow_resetlogs_corruption=true" 2)startup force 3) alter database open resetlogs. Dont' know whether i met your ans. :) Thanks for the video.
I prefer the data pump for cloning the database when there is no any direct connectivity between servers.
Cloning : The objective is to transfer the data[Binaries,and DB backup] from Source to Target Server either real time using RMAN duplicate or offline. By taking the backup and transfer the backup from source to target and do the clone on target site. By any mean need to transfer the data[OH binaries, & db backup] .
Thanks, very much ,bro
I perform the cloning first take the RMAN backup we can take and use Duplicate database method on the Target server ..
I perform the cloning first take the RMAN backup we can take and use Duplicate database method on the Target server .. very simple Question one type data refresh
By using NAS file system we can clone. Take the RMAN Backup into NAS mount point and then mount the same NAS file system on target and then start your cloning which means restore and recover.
I would say backup based cloning in case of network failure. But challenge is to move the high sized backup manually.
I just found this channel and so far it is great. Very informative.
Hi Arun excellent explanation .. awesome
Glad you liked it
@@dbagenesis I didn't thought that you will see my comment because its year ago post :) thank you so much..
then I want to share my Issue I faced with gap ..I had ODA standby db with ASM storage ..+ DATA was full then not applying any new archive from primary db .. I asked oracle support but they reply that the only solution is reinstall new ODA because +DATA is engineered to be one size we cant increase it .. it was not acceptable for my managers because it need more time and new budget .. so I solved it by rename datafiles from +DATA to +RECO because we have free size in +RECO ..then it worked successfully and standby db start to apply new archive archive again . I hope it help our dba community :)
Nice
Reg scan ip ---local lsnrs ..
Then what is the use of node vip ?? If scan lsnr is able to connect directly to node lsnr it can use the public IP directly right
Sir please try to talk about AutoUpgrade tool.
Thanks
and an additonal info remote_listener parameter wil be responsible for connection and current load balance
Can you please share one complete show to explain the concepts of VIP , SCAN IP , Scan listener , Remote Listener , HAIP and these are most confusing terminology in RAC and need some extremely deep understanding .. so please share your openion
Noted, this should be a Special Show! will work on it soon!
Hi, Could you please explain what changes made internally in oracle for which in active dataguard we can keep the stand by in read only with apply mode simultaneously . what changes made in the oracle database internally .
When you say "changes made internally" do you mean the exact queries to convert physical standby into active data guard?
@@dbagenesis I am extremely sorry for the late reply..the interviewer asked me same way ..I asked many guys but they did not get the answer and finally I got the answer from one of my friends working with Oracle corp. He told the answer would be _query_on_physical hidden parameter . we set it to true or false to activate or deactivate the active dataguard and seems oracle does not recommend this to do . Could you please cover up your idea on this on your next episodes and it will be helpful. I actually preparing for interview during those days and I was regularly watching your videos and it helped me a lot. thank you so much for such initiatives .
If catalog database is corrupted, how can rman search information
Can we get the row number of a selected record? If yes please let me know.I have inserted 1000 records in one table and I am trying to get the 999 id , and the rownumber and id should be same. when I try this always getting rownum is 1.
Basically I want to select Nth row from the table and it should match with the Id what I was inserted.
what query are you using?
@@dbagenesis I am using "SELECT rownum,id FROM emp a
GROUP BY rownum,id HAVING rownum = 10;"" and "select *from (select id, row_number() over (order by id desc) rnk from emp) where id=317"..but for small scale its working and if I have 2 Million rows both Queries taking time. and some times its not working.
What is transform parameter in expdp
Can you please explain about types of protection modes in dataguard and their purpose? How can I change the protection mode?
Matheen Ahamed DG setup relies on 3 protection modes,
Maximum Protection: Primary generates a redo which eventually will be received by standby and until acknowledgment comes from standby Primary won't commit also if there is a lag or network conn issue be primary and standby , Primary goes down. This mode is not commonly used
Maximum Performance - Primary issues commit to the user even if there is a delay in redo reception by standby
Maximum Availability- most commonly used. To put it in simpler way , DG setup acts as max protection mode for sometime and when a threshold breaches, it switches and acts as max performance.
Simple way to change protection mode is via DG broker
DGMRL> edit configuration set protection mode as Maxavailability;
But one must know what n all the parameters / specifications needed for a DG setup before adjusting protection
Modes
Happy learning :) Hope this helps
@@nishanthajay that's great Nishanth :)
Hi ,
I d have a question which process is responsible for sharing the 8 nodes load information and hopefully scan listener and local listener are different
scan listener works based on scan vip and local listener works based on node vip and scan vip.
Could you please clarify the above.
Sure, will answer this in upcoming episodes ;)
DBA Genesis Thank you! 😊
Hi,
If an L1 backup is performed today and no archivelog backup does not exist after L1 backup completes, will DB open after restore is done?
Yes, you can always forcefully open the DB but first, let's drill down.
To recover a database, you do need either of the two:
1. Archive log backups via RMAN OR
2. Archive logs on the disk (even if they are not backed up)
When you fire RMAN backups, by default RMAN performs a log switch and archives the redo logs. So you need that one archivelog in order to recover the DB.
All in all, its not about archivelog backups, its about archive logs. If you have them on the disk (which are not backed-up by RMAN) still you can perform normal recovery.
Would suggest you to master all RMAN scenarios on a test server so that you understand recover better. Have a look at one of our advance RMAN courses: dbagenesis.com/p/oracle-advanced-rman-backup-recovery
Hi,
Recently I have faced a problem.Lots of script deployed in one of of our production database i last weekend.
After that many queries performance goes decreased.we have no table or indexes are in Stale stats.The queries are running in the execution plan as earlier and in execution plan no full table scan happening.We have not found any index recommendation from Oracle tuning advisory.Can you please discuss what should our next action item and how can we tune database level parameter?
We are facing this issue after last weekend deployment.
Enable user level trace and read the trace file via TKPROF utility and see where exactly oracle is spending time.
@@dbagenesis thank you so much....
can we have more then 3 scan ip
Answer for 4: As per my knowledge if there is no coonectivity between source and target severs. Just create dblink in source database using target db entries, now transfer the backup files using dblink.
Without connectivity, how will you create DBLINK? How will your tns entries work without connectivity?
How you will transfer backup using dblink?
@@anuragkumarpandey using dbms file transfer utility we transfer the file from source to destination. Below command will work.
BEGIN
DBMS_FILE_TRANSFER.PUT_FILE(
source_directory_object => 'DATA_PUMP_DIR',
source_file_name => 'sample.dmp',
destination_directory_object => 'DATA_PUMP_DIR',
destination_file_name => 'sample.dmp',
destination_database => 'dblinkname'
);
END;
/
We use to copy the backup dump file from onpremise to rds and rds to rds severs in this way.
@@dbagenesis Correct. I have understand the quesyion wrongly.
Hi bro,
I have 2 questions for you.
1) how load balancing is archived in RAC.(is this round robin method means some sessions having long running some of sessions ends few seconds) how it's handle by Oracle.
2) odd number /even number of instance would you recommendation?why.
Brief explain active-passive scenario also...
A1. You need to read more about DNS, SCAN, SCAN listeners.
A2. Anything is fine, from 12c onwards the issue of ODD / EVEN is not a big deal.
A3. RAC is an active active cluster.
The all node load information stored in LBA(Load balancing advisory) so using this advisory scan listener assign connection to list loaded local listener.
if there's no connectivity between source and target db servers , then you cannot clone using RMAN.
How can we disable DML operation of end users, except from application server.
How about revoking specific permissions? example:
SQL> revoke update on hr.employees from scott;
SQL> revoke delete on hr.employees from scott;
SQL> revoke insert on hr.employees from scott;
Hi,
We don't want to revoke the privileges. i think i conveyed the requirement incorrectly.
My concern is that, we want our end users
to connect to the database only through our applications and then perform any dml.
The same user , if he connect to database using
SQL*PLUS or Microsoft acess using ODBC, or other third party tool, should not be allowed to perform DML.
@@tanwirrizwi1805 I don't think DB level we can restrict. I am assuming that how about Firewall settings. We will allow only application servers under that firewall.
@@tanwirrizwi1805 u can create a trigger and give the rights to only app server up to connect to dB server
Where are you man? Why you are not posting any more video?
Hi Arun sir, Can we have more than three scan listeners in scan configuration ?
Let's take even if you can configure more than three listeners, whats your point? Why would you need more than 3?
@Arun - You didnt answer this !
4. I want to clone database from source to Target but there is no connectivity between servers. How will clone happen with out connectivity and could you please share high level steps?
:) :) :)
Ok, so here are the high level steps:
1. Trigger backup on source
2. copy the backup pieces to external disk (USB)
3. On target server create same directory structure on new server
4. Register backup pieces with rman catalog command
5. Restore the database!
@@dbagenesis so basically .. this is db restore instead of cloning. Am I correct?
What will happen if undo ts if full?
What do you think what will happen? ;)
I think db will unstable
I don’t see your answer to the question you asked.