We have a process using Oracle AQ on Oracle 11.2. It was working just fine, buy yesterday, messages in the queue stopped getting dispatched to the subscriber. Here is the setup we have to create the queue and register the handler.
DBMS_AQADM.CREATE_QUEUE_TABLE (
queue_table => 'mdms_queue_table',
queue_payload_type => 'U$_MDM_QUEUE_OBJECT',
multiple_consumers => TRUE
);
DBMS_AQADM.CREATE_QUEUE (
queue_name => 'mdms_queue',
queue_table => 'mdms_queue_table'
);
DBMS_AQADM.START_QUEUE (
queue_name => 'mdms_queue'
);
DBMS_AQADM.ADD_SUBSCRIBER (
queue_name => 'mdms_queue',
subscriber => SYS.AQ$_AGENT(
'mdms_queue_subscriber',
NULL,
NULL )
);
DBMS_AQ.REGISTER (
SYS.AQ$_REG_INFO_LIST(
SYS.AQ$_REG_INFO(
'mdms_queue:mdms_queue_subscriber',
DBMS_AQ.NAMESPACE_AQ,
'plsql://P_MDMS_QUEUE_CB_PROCEDURE?PR=0',
HEXTORAW('FF')
)
),
1
);
DBMS_AQADM.GRANT_QUEUE_PRIVILEGE (
'ALL','UIMSMGR.MDMS_QUEUE', 'BAN_DEFAULT_M');
So now, when messages are posted to the queue, I run this query:
select DBA_QUEUES.NAME , aq.*
from gv$aq aq,
dba_queues
where aq.qid = dba_queues.qid
and
name like '%MDMS_QUEUE%'
And all the messages are sitting in the READY state. When I query v$process, I see that the program, "QMNC" is running as per this Oracle Troublshooting Doc. I can manually de-queue the messages by calling DBMS_AQ.dequeue(). Any suggestions on how to troubleshoot would be appreciated!
see this link for some aspects to be considered in oracle aq. Seems that some versions of the database have problems reported in metalink. Also try to recreate the objects if you have a working version before.
Related
Please excuse my ignorance as I am a total beginner.
I have created a basic RESTful Web Services Using PL/SQL, sample code below(which works perfectly fine)
ords.define_template(
p_module_name => 'restDemo',
p_pattern => 'updateClient');
ords.define_handler(
p_module_name => 'restDemo',
p_pattern => 'updateClient',
p_method => 'POST',
p_source_type => ords.source_type_plsql,
p_source => 'declare
w_clob clob := :body_text;
begin
ws.ws_interface.update_client(w_clob);
exception
when others then
package.ws_interface.g_result_rec.result_message := SQLERRM;
end;');
When ws_interface.update_client is initiated via the post request, in the initialization of ws_interface package, I am calling a routine thats determines if the client can be updated or not (based on access of the user logged in to the DB).
If the user does not have access to execute the updateClient service - I want to terminate the service call, which I have no idea on how to go about do this.
So my question is:
How do I go about terminating the current active updateClient webservice call ?
In your ws_interface package, if the client can not be updated, then you just need to return so that the PL/SQL block in the POST definition will complete. Your code can look something like this:
---Beginning of package
...
procedure update_client (some_clob clob) is
begin
--The client can not be updated so do no more processing.
if can_update_client = false then
--Some sort of error message could also be added here
return;
end if;
--Rest of processing for updating the client
end;
...
---End of package
Env: Oracle 12c
I currently have an Oracle Advanced Queue system setup as follows:
NAME QUEUE_TABLE QID QUEUE_TYPE MAX_RETRIES RETRY_DELAY RETENTION
--------------- --------------- ------- --------------- ----------- ----------- ---------
MY_WORK_Q MY_WORK_QT 2518333 NORMAL_QUEUE 100 0 0
MY_WORK_QT_E MY_WORK_QT 2518332 EXCEPTION_QUEUE 0 0 0
I also have registered a callback that calls a PL/SQL package procedure.
For some reason, I seem to have a situation where I am losing messages or messages are not being dequeued.
Based on this, I have the following questions:
Have I setup my actual queue - MY_WORK_Q with MAX_RETRIES and other info correctly?
Is there a way to check messages that have been enqueued?
Is there a way to check messages that have been de-queued?
Is there a means of checking the EXCEPTION_QUEUE/Exception table to see if lost messages have reached there?
I just can't see why I am losing messages within my queueing system and what I could check to see what might be causing the issue.
I also increased my MAX_RETRIES to 100 but still seem to be having issues.
Update
It seems like I am getting the error ORA-25263: no message in queue.
I am not sure if this is related to a timing issue on the dbms_aq.enqueue or dbms_aq.dequeue calls but on the dbms_aq.dequeue I have this set:
l_dequeue_options.wait := dbms_aq.no_wait;
Do I require say a 10 second wait instead of no_wait? I am unsure if this possible and whether the wait needs to be on the enqueue or dequeue steps.
Here is a simple example that works. Perhaps you can use it to identify your issue?
-- DROP TYPE some_type_t FORCE;
CREATE OR REPLACE TYPE some_type_t AS OBJECT (
a NUMBER,
b VARCHAR(100),
c DATE
);
CREATE OR REPLACE PROCEDURE some_type_callback(CONTEXT IN RAW,
reginfo IN sys.aq$_reg_info,
descr IN sys.aq$_descriptor,
payload IN RAW,
payloadl IN NUMBER) AS
-- Local variables
v_dequeue_options dbms_aq.dequeue_options_t;
v_message_properties dbms_aq.message_properties_t;
v_message_handle RAW(26);
v_some_type some_type_t;
BEGIN
-- Set the dequeue options from the descriptor
v_dequeue_options.consumer_name := descr.consumer_name;
v_dequeue_options.msgid := descr.msg_id;
-- Dequeue the message
dbms_aq.dequeue(queue_name => descr.queue_name,
dequeue_options => v_dequeue_options,
message_properties => v_message_properties,
payload => v_some_type,
msgid => v_message_handle);
END some_type_callback;
/
SELECT *
FROM user_errors e
WHERE e.name = 'SOME_TYPE_CALLBACK';
BEGIN
-- dbms_aqadm.drop_queue_table(queue_table => 'some_type_qt',
-- force => TRUE);
dbms_aqadm.create_queue_table(queue_table => 'some_type_qt',
queue_payload_type => 'some_type_t',
multiple_consumers => TRUE);
dbms_aqadm.create_queue(queue_name => 'some_type_q',
queue_table => 'some_type_qt',
retention_time => 86400); -- 1 day
dbms_aqadm.start_queue(queue_name => 'some_type_q');
dbms_aqadm.add_subscriber(queue_name => 'some_type_q',
subscriber => sys.aq$_agent(NAME => 'some_type_qs',
address => NULL,
protocol => NULL));
dbms_aq.register(sys.aq$_reg_info_list(sys.aq$_reg_info('some_type_q:some_type_qs',
dbms_aq.namespace_aq,
'plsql://some_type_callback',
hextoraw('FF'))),
1);
END;
/
SELECT *
FROM aq$some_type_qt;
-- nothing
DECLARE
v_some_type some_type_t;
eopt dbms_aq.enqueue_options_t;
mprop dbms_aq.message_properties_t;
enq_msgid RAW(16);
BEGIN
v_some_type := some_type_t(a => 42,
b => 'forty-two',
c => to_date('1/1/2942',
'mm/dd/yyyy'));
dbms_aq.enqueue(queue_name => 'some_type_q',
enqueue_options => eopt,
message_properties => mprop,
payload => v_some_type,
msgid => enq_msgid);
END;
/
SELECT *
FROM aq$some_type_qt;
-- msg_state = READY => PROCESSED
One thing that will definitely help you is setting retention_time on your queue table. You can then use the aq$ reference for the queue table to look at the messages. The msg_state column will show READY for a fresh message and PROCESSED for a message that was consumed. There are some other columns on there that can be helpful: retry_count for example.
If you are getting ORA-25263 it seems instead of handling the one message you are getting provoked for in your callback that you are trying to read a different message and clashing with another job that was started to consume the queue. This is resolved by the two lines I have in my callback before calling dequeue.
If you need to trigger when a message arrives and then preserve message order you need to add some locking and additional complexity to your callback. You could take a look at Metalink 225810.1 for examples on how to do this.
When i open new ticket in OS Ticket System, after ticket submission dump the array in the top of the page.
Error shown below.
Click to show image
Update
Error in text form:
Array ( [CSRFToken] => a56c08a9f49842de2ab3fc8e02d13ca2921d5f49 [a] => open [695fbbb799cdbbc2] => usman.saif22#gmail.com [ecf6c52d42557ef0] => M Usman [fec356fce5ea2de9] => [fec356fce5ea2de9-ext] => [topicId] => 2 [83163d48527225db] => Testing [message] => Testing [draft_id] => [emailId] => 0 [deptId] => 0 [cannedattachments] => Array ( ) )
This error occur when open new ticket, even ticket is successfully opened. But, shown this error when user open ticket.
How to dequeue messages from exception queue?
I searched in google but did not get any idea in brief.
First of all, you must start your exception queue:
EXECUTE DBMS_AQADM.START_QUEUE('OWNER.your_ex_q', true, true);
Then you can run the following instruction to dequeue, passing the MSG_ID:
DECLARE
dequeue_options dbms_aq.dequeue_options_t;
message_properties dbms_aq.message_properties_t;
message_handle RAW(16);
message aq.message_typ;
BEGIN
DBMS_AQ.DEQUEUE(queue_name => 'msg_queue',
dequeue_options => dequeue_options,
message_properties => message_properties,
payload => message,
msgid => message_handle);
COMMIT;
END;
If that's about Oracle Advanced Queuing (which I don't know), Google knows that Oracle says the following:
How do I dequeue from an exception queue?
The exception queue for a multiconsumer queue must also be a
multiconsumer queue.
Expired messages in multiconsumer queues cannot be dequeued by the
intended recipients of the message. However, they can be dequeued in
the REMOVE mode once (only once) using a NULL consumer name in dequeue
options. Messages can also be dequeued from an exception queue by
specifying the message ID.
Expired messages can be dequeued only by specifying message ID if the
multiconsumer exception queue was created in a queue table without the
compatible parameter or with the compatible parameter set to '8.0'
Is this what you're looking for?
Firefox and Chrome report ~2000ms response time (localhost) to my REST API. The database executes the query in under 1ms. A static JSON file with the same content is served in 4ms.
Where is all the time going? Is ORDS.war consuming it, is it database connection time, is Tomcat using it or something else? How do I find the root cause?
Update: Win10. Stopped Apache service and tried running catalina.bat from command line and response time drops to ~30ms. Killed it and tried running as a service again - response time consistently ~170ms.
I have an ORDS 3.0.11.180.12.34 service running on Tomcat 8.5.16 (jre1.8.0_144 x64) running this query:
SELECT 1 FROM dual
as below
DECLARE
l_handler_id ords_metadata.ords_handlers.id%TYPE := NULL;
l_param_id ords_metadata.ords_parameters.id%TYPE := NULL;
l_module_name ords_metadata.ords_modules.name%type := 'rest';
l_pattern ords_metadata.ords_templates.uri_template%type := 'select_1_from_dual';
l_method ords_metadata.ords_handlers.method%type := 'GET';
BEGIN
ords.define_template (
p_module_name => l_module_name,
p_etag_type => 'QUERY',
p_etag_query => q'~SELECT dbms_random.value(0,1) FROM dual~',
p_pattern => l_pattern
);
ords.define_handler (
p_module_name => l_module_name,
p_pattern => l_pattern,
p_method => l_method,
p_source_type => ords.source_type_collection_item,
p_source => 'SELECT 1 FROM dual'
);
COMMIT;
END;
/