Select query over database link hanging

Hi All,
I have been looking all over the internet and all over this forum but i could not find the solution for my issue given below.
There are a number of distributed databases (Oracle 10g) over a private network. Sending and receiving data amongst all the databases work fine. But when one particular database A tries to receive data from database B, it hangs and then the famous old error TNS LOST CONTACT.
Basically, this has been implemented by creating a batchjob. This worked fine for years and now suddenly this is not working only for the two specific databases.
It is just a simple select query being hit on the remote database. This is called through a function in the package, which builds the dynamic select query using DBMS_SQL package.
It actually hangs when doing DBMS_SQL.PARSE ();
The same function when called in a pl/sql block from database A, works perfectly fine and receives the required data from database B. This was tried with Oracle SQLDeveloper, TOAD and SQLPLUS.
Things seems to be working fine, but actually unaware of the problem. I tried all the suggestions provided over the metalink also.
Require your help urgently.
Thanks
Akhil.

Hi Rajat,
The user uses DEFAULT profile. The details of the parameter are:
PROFILE RESOURCE_NAME RESOURCE_TYPE LIMIT
DEFAULT COMPOSITE_LIMIT KERNEL UNLIMITED
DEFAULT SESSIONS_PER_USER KERNEL UNLIMITED
DEFAULT CPU_PER_SESSION KERNEL UNLIMITED
DEFAULT CPU_PER_CALL KERNEL UNLIMITED
DEFAULT LOGICAL_READS_PER_SESSION KERNEL UNLIMITED
DEFAULT LOGICAL_READS_PER_CALL KERNEL UNLIMITED
DEFAULT IDLE_TIME KERNEL UNLIMITED
DEFAULT CONNECT_TIME KERNEL UNLIMITED
DEFAULT PRIVATE_SGA KERNEL UNLIMITED
DEFAULT FAILED_LOGIN_ATTEMPTS PASSWORD 10
DEFAULT PASSWORD_LIFE_TIME PASSWORD UNLIMITED
DEFAULT PASSWORD_REUSE_TIME PASSWORD UNLIMITED
DEFAULT PASSWORD_REUSE_MAX PASSWORD UNLIMITED
DEFAULT PASSWORD_VERIFY_FUNCTION PASSWORD NULL
DEFAULT PASSWORD_LOCK_TIME PASSWORD UNLIMITED
DEFAULT PASSWORD_GRACE_TIME PASSWORD UNLIMITED
Thanks
Akhil

Similar Messages

Slow Query over Database Link with Bind Variable

I have a query over a DB link, with all tables on the remote database.
If I use a bind variable (from Toad), the query takes 4 minutes. If I replace the bind variable with a constant or substitution variable, it takes 1 second.
The query runs fine when run directly on the remote database using bind variable.
9.2.0.7

Look up "Bind variable peeking"
What's happened is you have an execution plan that differs from the one with the constant. Why? My bet is that Oracle "peeked" at the bind variable to help it decide which execution plan to build. It then cached it. It probably cached an execution with an index when it should be doing a full table scan or a hash join instead of a nested loop. It's hard to say specifically what it is.
Try this, flush your shared pool and rerun the query with the bind and let us know if it takes 1 second or 4 minutes. If it takes 1 second, then that was probably it.
Read part 2 of Tom Kyte's blog post on what it is and it's behavior.
http://tkyte.blogspot.com/2007/09/sqltracetrue-part-two.html

Interactive report performance problem over database link - Oracle Gateway

Hello all;
This is regarding a thread Interactive report performance problem over database link that was posted by Samo.
The issue that I am facing is when I use Oracle function like (apex_item.check_box) the query slow down by 45 seconds.
query like this: (due to sensitivity issue, I can not disclose real table name)
SELECT apex_item.checkbox(1,b.col3)
, a.col1
, a.col2
FROM table_one a
, table_two b
WHERE a.col3 = 12345
AND a.col4 = 100
AND b.col5 = a.col5
table_one and table_two are remote tables (non-oracle) which are connected using Oracle Gateway.
Now if I run above queries without apex_item.checkbox function the query return or response is less than a second but if I have apex_item.checkbox then the query run more than 30 seconds. I have resolved the issues by creating a collection but it’s not a good practice.
I would like to get ideas from people how to resolve or speed-up the query?
Any idea how to use sub-factoring for the above scenario? Or others method (creating view or materialized view are not an option).
Thank you.
Shaun S.

Hi Shaun
Okay, I have a million questions (could you tell me if both tables are from the same remote source, it looks like they're possibly not?), but let's just try some things first.
By now you should understand the idea of what I termed 'sub-factoring' in a previous post. This is to do with using the WITH blah AS (SELECT... syntax. Now in most circumstances this 'materialises' the results of the inner select statement. This means that we 'get' the results then do something with them afterwards. It's a handy trick when dealing with remote sites as sometimes you want the remote database to do the work. The reason that I ask you to use the MATERIALIZE hint for testing is just to force this, in 99.99% of cases this can be removed later. Using the WITH statement is also handled differently to inline view like SELECT * FROM (SELECT... but the same result can be mimicked with a NO_MERGE hint.
Looking at your case I would be interested to see what the explain plan and results would be for something like the following two statements (sorry - you're going have to check them, it's late!)
WITH a AS
(SELECT /*+ MATERIALIZE */ *
FROM table_one),
b AS
(SELECT /*+ MATERIALIZE */ *
FROM table_two),
sourceqry AS
(SELECT b.col3 x
       , a.col1 y
       , a.col2 z
FROM table_one a
    , table_two b
WHERE a.col3 = 12345
AND   a.col4 = 100
AND   b.col5 = a.col5)
SELECT apex_item.checkbox(1,x), y , z
FROM sourceqry
WITH a AS
(SELECT /*+ MATERIALIZE */ *
FROM table_one),
b AS
(SELECT /*+ MATERIALIZE */ *
FROM table_two)
SELECT apex_item.checkbox(1,x), y , z
FROM table_one a
    , table_two b
WHERE a.col3 = 12345
AND   a.col4 = 100
AND   b.col5 = a.col5If the remote tables are at the same site, then you should have the same results. If they aren't you should get the same results but different to the original query.
We aren't being told the real cardinality of the inners select here so the explain plan is distorted (this is normal for queries on remote and especially non-oracle sites). This hinders tuning normally but I don't think this is your problem at all. How many distinct values do you normally get of the column aliased 'x' and how many rows are normally returned in total? Also how are you testing response times, in APEX, SQL Developer, Toad SQLplus etc?
Sorry for all the questions but it helps to answer the question, if I can.
Cheers
Ben
http://www.munkyben.wordpress.com
Don't forget to mark replies helpful or correct ;)

Report over database link error

Good morning all.
I have a database link LINKA from USERA on LOCALSERVER to USERB on REMOTESERVER.
(create database link LINKA connect to USERB identified by mypassword using validtns)
In USERC on REMOTESERVER I have a TABLEX, with grant all on TABLEX to USERB on REMOTESERVER.
In USERB on REMOTESERVER I have a view TABLEXV, select * from USERC.TABLEX
From USERA on LOCALSERVER I can select * from TABLEXV@LINKA. In fact I can insert/update/delete tablexv from USERA.
I also have a view TABLEXV on USERA, select * from TABLEXV@LINKA. I can also select/insert/update/delete from USERA.TABLEXV.
So far so good - all working as expected.
Then, my APEX app on LOCALSERVER, interactive report, parsing schema USERA, querying TABLEXV, I get the following:
when "Exclude Link Column", I get the error "The requested URL /pls/apex/f was not found on server.
when "Link to Single Row View", report displays correctly.
when "Link to Custom Target", I get the error "The requested URL /pls/apex/f was not found on server.
When I run the select * from TABLEXV as a regular report (i.e. not interactive), I also get the error.
So the only time I get the report displayed correctly is using IR and "Link to Single Row View".
However, clicking on the link (for Single Row View) I get "The requested URL /pls/apex/wwv_flow.show was not found on this server.".
To summarize:
- accessing table via view over database link
- works outside of apex without issue
- inside only works when Link to Single Row View set
- But clicking on link fails
I could live with just having the Link to Singe Row View option if I could disable displaying of the icon so that the user cant click on it.
APEX 3.2.00.27
EE Database - 10.2.0.3 (yes I know out of support)
Remote EE Database - 10.2.0.5
Real object names replaced to protect the innocent!
Anyone got any clues?
I have used remote tables many times in my APEX applications and never had this issue.
Thanks,
Rob @ very confused .com

Good morning all.
I have a database link LINKA from USERA on LOCALSERVER to USERB on REMOTESERVER.
(create database link LINKA connect to USERB identified by mypassword using validtns)
In USERC on REMOTESERVER I have a TABLEX, with grant all on TABLEX to USERB on REMOTESERVER.
In USERB on REMOTESERVER I have a view TABLEXV, select * from USERC.TABLEX
From USERA on LOCALSERVER I can select * from TABLEXV@LINKA. In fact I can insert/update/delete tablexv from USERA.
I also have a view TABLEXV on USERA, select * from TABLEXV@LINKA. I can also select/insert/update/delete from USERA.TABLEXV.
So far so good - all working as expected.
Then, my APEX app on LOCALSERVER, interactive report, parsing schema USERA, querying TABLEXV, I get the following:
when "Exclude Link Column", I get the error "The requested URL /pls/apex/f was not found on server.
when "Link to Single Row View", report displays correctly.
when "Link to Custom Target", I get the error "The requested URL /pls/apex/f was not found on server.
When I run the select * from TABLEXV as a regular report (i.e. not interactive), I also get the error.
So the only time I get the report displayed correctly is using IR and "Link to Single Row View".
However, clicking on the link (for Single Row View) I get "The requested URL /pls/apex/wwv_flow.show was not found on this server.".
To summarize:
- accessing table via view over database link
- works outside of apex without issue
- inside only works when Link to Single Row View set
- But clicking on link fails
I could live with just having the Link to Singe Row View option if I could disable displaying of the icon so that the user cant click on it.
APEX 3.2.00.27
EE Database - 10.2.0.3 (yes I know out of support)
Remote EE Database - 10.2.0.5
Real object names replaced to protect the innocent!
Anyone got any clues?
I have used remote tables many times in my APEX applications and never had this issue.
Thanks,
Rob @ very confused .com

Problems with query via Database-Link (Oracle 7.3.4)

I made the following simple query via database-link.
select count (*) from [email protected]
I got the result in about 200 miliseconds.
Then I4ve tried the following query.
select * from [email protected]
It took about 2 hours end ended with an error message "ORA-03113: end-of-file on communication channel"
So I made another choice
select * from [email protected] where rownum <=1
Took about 200 miliseconds.
So I made the query again an changed only the number of rownums,until I was at rownum <=8
then it went sleeping again(I've canceled the statement because I didn't want to wait 2 hours again).
We made the same things from another server with the same database-link on the same remote server
and we had no problems.
Any Idea???

No, network is OK!
Other links are running, I have also no problems when I connect directly to the remote server.

Can we create pl/sql code over database link?

Hi All,
We are using Oracle 9i database.
I want to know if we can create or modify pl/sql code (procedue, package, etc) over the database link? That means - can we create a procedure in remote database using the db link?
Thanks,
Dnyanesh

yes, I can connect to the remote database directly and create/modify the packages.
But I want the users not to create them using the db link.
Is there any way to create the pl/sql package over database link?
Thanks,
Dnyanesh

ORA-01653 error while executing a select query over DBLINK

Hi,
We have a query that is running to extract some data from a remote DB over DBLINK.
The query is failing by throwing the error "ORA-01653 -- Unable to extend.........".
Is this a problem with the "temp" tablespace of local database or is this with the "temp" tablespace of remote DB.
As far as my knowledge goes, I guess any statement , selecting data over dblink where the query is being fired actually in remote db over dblink uses the temp tablespace of that DB only and hence the problem should be with the temp tablespace of remote db only but I am not very sure about this.
Is there any chance that it can be problem with "temp" tablespace of local DB from where the query is being fired over dblink.
It would be really helpful if anyone can throw some light on this.
Thanks

The error stack will normally tell you if the exception is raised in the remote database.
You might try using the driving_site hint (see performance guide) to push work to the remote site or pull it local. It all depends on your particular query and explain plan.
Remember to use { code } (without the embedded spaces) tags to frame your code and explain plan so it remains formatted, if you post it here.

Execute procedure over database link

Hi,
I''m working on a package where you can select the required database, paste some query and with clicking on a button it will execute and create an explain plan for this query.
Therefore i've create a procedure on each database that put's the explain plan in the PLAN_TABLE.
To do this i need to execute this prodecure from the package. I use the following code:
DBA_OWNER.DB_PCD_EXPLAIN_PLAN@DBA_LINK_533.WORLD(p_query);
This is working fine as this database link allready exists to the required database. But i want to make the database link name variable. I've allready created the dynamic database link:
l_link varchar2(20) := 'DBA_LINK';
l_link_nr number(4) := dbms_random.value(0,1000);
l_db varchar2(20) := l_link||'_'||l_link_nr;
l_statement varchar2(4000);
begin
l_statement := 'create database link '||l_db||' connect to **** identified by ***** using '''||p_database||'.WORLD''';
execute immediate l_statement ;
So now i have the dynamic databaselink that is created each time i call the package.
now i want to use this databaselink to call the procedure. this is where it goes wrong.
The code i'm using:
DBA_OWNER.DB_PCD_EXPLAIN_PLAN@l_db(p_query);
is not acceoted when i build the package.
The follwoing error message is showed.
PACKAGE BODY DBA_OWNER.PCK_EXPLAIN_PLAN
On line: 66
PLS-00352: Unable to access another database 'L_DB'
Does someone know how to use this variable to execute the procedure on the remote db?

Any procedures you call have to be accessible at compile time, so the database link must also exist at compile time.
You can get around this by putting your procedure call to DB_PCD_EXPLAIN_PLAN inside dynamic SQL, the same as you are doing to create the database link. Dynamic SQL is not checked until runtime.
Think about the risk if this procedure is exploited. You are creating a way for someone to run arbitrary code in an any database in your environment, with presumably a highly-privileged ID, if it can run an explain plan against any schema in any target database.
This is exposing some significant security issues. Read up on SQL injection, and have someone else review the code for security issues before deploying this.

Passing parameters to select query in DB link from invoke...

I am having a DB partner link which executes the select query based on user specified parameter.
Now i have created an invoke activity and specified input[this needs to be passed to query] and output parameters[this is query result]. This invoke activity calls DB partner link in which one parameter has been created to be used in select query. So my doubt is , will DB partner link take up the input parameter of invoke activity automatically or do we need to specify any mapping for this?

When you create a db partnerlink the parameter you create for your select is the input and when you create the input variable in the invoke for this partnerlink it is created with the right type for the parameter.
Create an Assign before your invoke and assign a value to the input variable you created in the invoke. That value will be passed and used as the parameter in the select.
Heidi.

Can clob columns be selected across a database link (synonym)?

Hello,
I would like to do something like this...
insert into archive_table (id, msg, start_date)
select id, msg, startdate from synonym_table
where start_date < whatever;
...where msg represents a clob column and synonym_table represents a synonym in the current table space which accesses another Oracle table using a database link.
Is this possible?

http://asktom.oracle.com/pls/ask/f?p=4950:8:15261108079829288196::NO::F4950_P8_DISPLAYID,F4950_P8_CRITERIA:950029833940

Database link hangs from source 10204 to remote 9208

Hi,
I have a sql job that runs sql statements on database 10204 connecting to database 9208 via a database link. The job takes about half hour to complete. The source database, 10204 connects to remote 9208 to retreive data from 9208. Most of the processing is performed on the remote db. For some reason, after processing sql statements on remote db, the job hangs. I looked at wait events on remote and it shows "SQL*Net message from client". I looked at wait events on source database and it shows "SQL*Net message from dblink".
It looks to me that there is something wrong with the communication between the source and remote db via the db link.
I tried setting the default_sdu_size from 2k to 8k but this does not help the hanging. I tried runing the job at different times in the day and night but this does not help either. I filed an SR with metalink and they asked me to run trace. I sent them the trace files and they came back saying to bounce the databases. I bounce the databases and still that did not resolve the hanging. Anyone encounter this issue lately? Thank you in advance.

I have seen this before between 2 10g databases, but very rarely, and never identified the cause, so suspected it was from a network glitch. Solution was to kill the hung sessions and restart.
Setup some network tests and leave them running. E.g ping tests, scp tests, and/or open ssh sessions to see if you lose any packets when the database link fails. Check your OS logs for network interface errors. Check if there are any firewalls between the databases which may disconnect inactive sessions.

Run process flow over database link

Hi guys,
I have two projects here, A and B. Each project has its own process flow, lets say Process A (in project A)and process B(in project B). I need to run process B in process A, so I use the database link, and WB_RT_API_EXEC.run_task function. But when I run the process, I get the following error:
ORA-20001: Task not found - Please check the Task Type, Name and Location are correct.
ORA-06512: at "OWBRTR2.WB_RT_API_EXEC", line 704
does anyone have some idea how to fix this? or any other suggestions for how to do it? thanks.

Hi Gowtham,
I am using the following function to send execution details of the process-flow. This process-flow is having 20 mappings. This is working fine.
The cIient requirment is to pass email-IDs from a back-end table(sam_mail_notification) with columns(smtp, from_address and to_address), right now the function is working on static(constant) values.
I had modified the function according to the requirement, can you please have a look at the function, and tell me is it correct. Because the values should come from database table. will this function work.
I got the following error, while deploying,
ORA-06550: line 55, column 1:
PLS-00103: Encountered the symbol "END" when expecting one of the following:
FUNCTION SEND_SUMMARY1()
RETURN NUMBER
--initialize variables here
-- main body
retval number := 0; --default
crlf CONSTANT VARCHAR2(2):= CHR(13) || CHR(10);
pSender VARCHAR2(30) := '[email protected]';
pRecipient VARCHAR2(30) := '[email protected]';
pSubject VARCHAR2(100) := 'Process REsults for: '||to_char(sysdate,'dd/mm/yyyy');
mesg VARCHAR2(32767);
mail_conn utl_smtp.connection;
cursor getResults is
select AREA.OBJECT_NAME,
AREA.CREATED_ON,
AREA.UPDATED_ON,
aramr.ELAPSE_TIME,
NUMBER_RECORDS_SELECTED,
NUMBER_RECORDS_INSERTED,
NUMBER_RECORDS_UPDATED,
NUMBER_ERRORS,
AREA.RETURN_RESULT,
AREA.EXECUTION_AUDIT_STATUS,
MESSAGE_SEVERITY,
MESSAGE_TEXT
from
all_Rt_audit_executions area,
all_Rt_audit_map_runs aramr,
all_rt_audit_exec_messages err
where AREA.execution_audit_id = ARAMR.execution_audit_id(+) AND
AREA.execution_audit_id = err.execution_audit_id(+)
and
trunc(area.created_on) = trunc(sysdate)
AND AREA.OBJECT_NAME IS NOT NULL AND AREA.TASK_TYPE!='ProcessFlow'
AND area.top_level_execution_audit_id =(select max(top_level_execution_audit_id)from all_Rt_audit_executions
where execution_name = (select execution_name from all_Rt_audit_executions where task_type ='ProcessFlow' and top_level_execution_audit_id =
(select max(top_level_execution_audit_id) from all_Rt_audit_executions)));
Cursor getMailDetails is select
mailhost
pSender,
pRecipient
from sam_mail_notification;
BEGIN
for maild in getMailDetails
loop
mail_conn := utl_smtp.open_connection(mailhost, 25);
mesg := 'Date: ' ||TO_CHAR( SYSDATE, 'dd Mon yy hh24:mi:ss') || crlf ||
'From: <'|| pSender ||'>' || crlf ||
'Subject: '|| pSubject || crlf ||
'To: '||pRecipient || crlf || '' || crlf ||
end loop;
'MAPNAME START_TIME END_TIME ELAPSED SELECTED INSERTED UPDATED ERRORS RESULT AUDIT_STSTUS SEVERITY MESSAGE_TEXT'||crlf ;
for rec in getResults
loop
mesg:=mesg || rpad(rec.OBJECT_NAME,30)||
rpad(to_char(rec.CREATED_ON, 'HH24:MI:SS DD-MON-YY'),20)||
rpad(to_char(rec.UPDATED_ON, 'HH24:MI:SS DD-MON-YY'),24)||
rpad(to_char(rec.ELAPSE_TIME),7)||
rpad(to_char(rec.NUMBER_RECORDS_SELECTED),8)||
rpad(to_char(rec.NUMBER_RECORDS_INSERTED),10)||
rpad(to_char(rec.NUMBER_RECORDS_UPDATED),10)||
rpad(to_char(rec.NUMBER_ERRORS),5)||
rpad(to_char(rec.RETURN_RESULT),8)||
rpad(to_char(rec.EXECUTION_AUDIT_STATUS),10)||
rpad(to_char(rec.MESSAGE_SEVERITY),10)||
rpad(to_char(rec.MESSAGE_TEXT),200)||crlf;
end loop;
utl_smtp.helo(mail_conn, mailhost);
utl_smtp.mail(mail_conn, pSender);
utl_smtp.rcpt(mail_conn, pRecipient);
utl_smtp.data(mail_conn, mesg);
utl_smtp.quit(mail_conn);
return retval;
EXCEPTION
WHEN OTHERS THEN
return 1;
END;
Regards,
Kumar

Column reference over database link.

Hi,
Is it possible to reference a table column across a database link?
Ex: var1 table.column%TYPE@dblink
Thanks
Chad

Chad,
Glad it worked ..
Taking this thread a step further ..
If you have the privileges to create synonyms then you could do the following:
create synonym mySynonym
for table@dblink
Then use the synonym in the PL/SQL code as follows:
var1 mySynonym.column%TYPE;
This will avoid repetitive typing of table.column@dblink and if the dblink name changes then you have to make the change in only one place - the synonym
Shakti
http://www.impact-sol.com
Developers of Guggi Oracle - Tool for DBAs and Developers

Query on database links

I wanted to know what is the difference between these two types of database links:
create public database link <>
connect to <>
identified by <>
using 'XYZ';
and
create public database link <>
connect to <>
identified by <>
using '(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = 1.1.7.1)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SID = XYZ)
For me the second one works but at some point it started giving me errors
ORA-02020, till I hit commit.
Dhiren

I would expect ORA-02020 to occur with other method also as it has nothing to do with how the link is created. It all depends on how many connections are open at a time.
Check the limit imposed by open_links parameter.

Interactive report performance problem over database link

Hi gurus,
I have an interactive report that retrieves values from two joined tables both placed on the same remote database. It takes 45 seconds to populate (refresh) a page after issuing a search. If I catch the actual select that is generated by apex ( the one with count(*) over ()...) and run it from sql environment, it takes 1 or 2 seconds. What on earth does consume the rest of the time and how to speed this up? I would like to awoid creating and maintaining local materialized views if possible.
Regards
Samo

Hi
APEX normally needs to return the full result set for the purposes of pagination (where you have it set to something along the lines of show x-y of z), changing this or the max row count can affect performance greatly.
The driving site hint would need to refer to the name of the view, but can be a bit temperamental with this kind of thing. The materialize hint only works for sub-factored queries (in a 'WITH blah AS(SELECT /*+ MATERIALIZE */ * FROM etc. etc.)). They tend to materialize anyway and its best not to use hints like that for production code unless there is absolutely no other option, but just sub factoring without hints can often have a profound effect on performance. For instance
WITH a AS
SELECT c1,
        c2,
        c3,
        c4,
        c5
FROM schema1.view1)
, b AS
SELECT c1,
        c2,
        c3
FROM schema1.view2)
SELECT *
FROM a, b
WHERE a.c5 = b.c3May produce a different plan or even just sub factoring one of the external tables may have an effect.
You need to try things out here and experiment and keep looking at the execution plans. You should also change Toads row fetch setting to be all 9's to get a real idea of performance.
Let me know how you get on.
Cheers
Ben
http://www.munkyben.wordpress.com
Don't forget to mark replies helpful or correct ;)
Edited by: Munky on Sep 29, 2009 1:41 PM

Select query over database link hanging

Similar Messages

Maybe you are looking for