Question on v$ session.status = KILLED

Similar Questions

• Why the DBMS_JOBs restart after the session was killed?

  DB version: 10.2.0.4
  Platform: Solaris 5.10
  
  We were running a via DBMS_JOB procedure. We wanted to cancel execution of this job. So I found the session whose JOB it was running using the column v$ session.program (J000 chain there). Then I performed a

  ALTER SYSTEM KILL SESSION '889,3354';

  The session was killed. But after a few moments, this work has got automatically restarted!
  
  Question1
  Is this (restart) after kill session the default behavior of DBMS_JOB?
  
  Question2
  Is this (restart) after kill session DBMS_SCHEDULER default behavior?
  
  Question3.
  How to kill a DBMS_JOB who is running. Surprisingly, I couldn't find this in the documentation
  
  http://download.Oracle.com/docs/CD/B19306_01/AppDev.102/b14258/d_job.htm#i1000521
  
  For DBMS_JOB. BROKEN the clause, that's what says the documentation
  
  If you set work as incorrect while it is running, Oracle resets the situation of normal employment once the work is completed. Therefore, perform this procedure for jobs that are not running.

  Q1:

  Yes, when a task is run in an error it will retry again 16 times (9i) until it sets itself to 'broken '.

  http://asktom.Oracle.com/pls/Apex/f?p=100:11:0:P11_QUESTION_ID:2328991547163 #1750919700346652272

  http://download.Oracle.com/docs/CD/B19306_01/server.102/b14231/schedadmin.htm#sthref3798

  You can query ALL | DBA_JOBS for more information on how your work ran, including the number of FAILURES;
  http://download.Oracle.com/docs/CD/B19306_01/server.102/b14237/statviews_1086.htm#i1580430

• Question on V$ SESSION. STATUS column

  DB version: 11.2.0.4 on Oracle Linux 6.4

  V$ SESSION. STATUS: My question is about the IDLE State.

  Under what circumstances does V$ SESSION. STATUS become INACTIVE for a session?

  For a particular run SQL, if the session has to wait a/o lock or Table being released, will be V$ SESSION. Change of STATUS to "INACTIVE"?

  If the session is waiting for an i/o for a lock or a lock, its status is ACTIVE.

  However, after the current call ends its status is INACTIVE.

  If the session sends multiple lines to a customer and the number of lines, such that it exceeds the size of the array, the session must make several calls to send the rows to the client.  Between calls, even if she is waiting for the client to recognize the last set of rows sent, its status may be INACTIVE.

  Similarly, while the session waits for the next SQL statement of the client, its status would be INACTIVE.

  Hemant K Collette

• Why DB read file is still active after that the session was killed?

  Hello
  
  I have a query which is only select read only (no selection for update) on a big enough table (lines 500millions). The result must return a few million rows (1-10 million), because it is not properly written.
  
  I killed all the sessions opened by the user who started the application, but after 3 hours, it is still ongoing.
  
  I don't understand. It's a read. There is no data restoration that must be set aside, there is no generation of redo (low) because of this.
  There are a few kinds made in this query.
  
  So, why after that I'm killing a query that is read-only, the process runs for a very long time?
  
  It is not the first time I'm facing this problem.
  
  Any ideas?
  
  Thank you
  Daniel

  >
  I killed all the sessions opened by the user who started the application, but after 3 hours, it is still ongoing.
  >
  Have Yes - but you used the IMMEDIATE clause to kill HER?
  >
  IMMEDIATE specify IMMEDIATE load the Oracle database to ongoing transactions, releases all session locks, recover session state full and immediately returns control to you.
  >
  If you really want the process completed rather than simply "marked" termination then use DISCONNECT SESSION with the IMMEDIATE clause

  See the "Clause of DISCONNECT SESSION" in the doc of the SQL language
  http://docs.Oracle.com/CD/B28359_01/server.111/b28286/statements_2013.htm
  >
  Use the DISCONNECT SESSION clause to disconnect the current session to destroy the dedicated server process (or virtual circuit if the connection was made through a shared server).
  . . .
  The IMMEDIATE environment disconnects the session and retrieves the status of entire session immediately, without waiting for open transactions to complete.
  >
  View the results of the use of DISCONNECTION

• Question about discovery/inventory/status schedule

  Hello.

  I have a question.

  Can I disable schedule discovery/inventory?

  I'll manually run discovery/inventory during the installation of new devices.

  and then I turn off hourly status?

  If the calendar of the disabled state, draw the attention of job function properly?

  Hi and thanks for the post.

  You can go to the Discovery Portal page - the place that you have created your discovery beaches... from there, you should find an arrangement to clear the calendar of discovery/inventory.

  With regard to the status, I would like to keep it to 1 hour.  This is the survey that will check the health of hardware for all your devices.

  If, however, you use only OME as SNMP trap receiver, you can disable this too I guess.  But the icon of health on the tree view of the devices will probably not updated unless a trap happens on a particular device.

  Thank you

  Rob

  delltechcenter.com/OME

• kill session vs kill session immediate/post_transactional

  A DBA, I sometimes kill sessions in SQL * more by entering the command ' ALTER SYSTEM KILL SESSION ' xxx, xxxx '; The session ends quickly, sometimes it takes sometimes minutes for the killed session to go.
  
  When to use Enterprise Manager Grid Control to kill a session that I have to choose between a 'kill immediately' or 'transactional Message. When I kill a session by entering a command from SQL * more I do not use one of these options.
  
  I'm trying to understand the consequences of the use of the two options available via the grid control. Clause of some forum posts suggest you use IMMEDIATELY"as a last resort. According to them, he couldn't prevent a session from filling out a cancellation of the updates because the operating system process stops immediately. Read that a description of the clause POST_TRANSACTION makes me think of issueing what the command against a long query term will allow him continue to run and end the session only after the end of the SQL query. For now I avoid using one of the options of the grid control and stick to my original format in SQL * more.
  
  Could someone shed some light on the difference between killing a session with or without KILL it IMMEDIATELY or POST_TRANSACTIONAL' options?
  
  Thank you
  Bill

  Hello

  If I order, "the ALTER SYSTEM KILL SESSION ' xxx, xxxx ';

  allows you to mark a session as null and void, restore the ongoing transactions, releases all session locks and partially recovery of resources from the session. If the session is running an activity that must be completed, as the wait for a response from a remote database or rolling back a transaction, then Oracle Database expects this activity complete, mark the session as being finished and then returns control to you.

  If I order, "ALTER SYSTEM KILL SESSION ' xxx, xxxx" IMMEDIATE. "

  Specify IMMEDIATE to load the Oracle database to ongoing transactions, releases all session locks, recover session state full and immediately returns control to you. You wrote end immediately with no cancellation of updates?. This is not so, it will rollback all not updated.

  If I execute the command ' ALTER SYSTEM KILL SESSION ' xxx, xxxx "POST_TRANSACTION;"

  The POST_TRANSACTION parameter allows operations to be completed before the session is disconnected. If the session has no current transaction, then this clause has the same effect as to KILL SESSION.

  Reference:-[http://download.oracle.com/docs/cd/B19306_01/server.102/b14200/statements_2013.htm#SQLRF00902]

  HTH
  Anand

• Question 3G indicator of status.

  This has probably been asked a million times, and I'm sorry if there. But I have a question about the 3G symbol. It has two arrows below the top and down, when they are in gray, that means the phone is not connected? or the fact that the 3G symbol is it mean its connected? If this is the case how to disconnect and let him connect automatically and only when I ask?
  I did a little research on the internet and all I get is that it is an indicator of the 3G and its 3 faster then E. Which is not useful at all. I'm worried about her be connected all the time - course fees - and if there is anyway to control.

  hex90 wrote:If you want to switch between 3G and 2G to save battery: Settings | Wireless Controls | Mobile Networks | Select Network.If you want to turn off data w/out using airplane mode that also turns off your calling (which it sounds like), search the Market for APNdroid. Excellent app that will do this. It's lightweight. Just make sure you have it disabled (i.e. have Internet access) before uninstalling. It doesn't do anything with wifi.
  

• Question of the PAP status query

  I have a message APP and you want to query the status using .  However, when I send the message, the MDS Simulator gives an error "incorrect message of PAP: bid invalid push". ""  I don't know what the deal is.  Does anyone have an idea what is the problem?  Here's a little piece of my code making the call.  Of course, I actually use a push-valid identification and address_value.

  ....
  
  HttpURLConnection MDS_CONN = (HttpURLConnection) MDS_HOST_URL.openConnection();
  
  MDS_CONN.setRequestProperty("Content-Type", "multipart/related; type=\"application/xml\"; boundary=" + boundary);
  MDS_CONN.setRequestProperty("X-Wap-Application-Id", "/");
  MDS_CONN.setRequestProperty("X-Rim-Push-Dest-Port", DEVICE_PORT);
  MDS_CONN.setRequestMethod("POST");
  
  MDS_CONN.setAllowUserInteraction(false);
              MDS_CONN.setDoInput(true);
              MDS_CONN.setDoOutput(true);
  
  String Message = "";
  Message += "Content-Type: application/xml\r\n\r\n";
  Message += "\r\n";
  Message += "http://www.wapforum.org/DTD/pap_2.0.dtd\">\r\n";
  Message += "\r\n";
  Message += "\r\n";
  Message += "
  \r\n";
  Message += "\r\n";
  Message += "\r\n";
  
  ....
  
  This is what the message looks like:
  
  Content-Type: application/xml
  
  
  http://www.wapforum.org/DTD/pap_2.0.dtd">
  
  
  
  
  
  

  Thanks, I've thought about it.  Here's how it should look like instead of my original code I posted.

  HttpURLConnection MDS_CONN = (HttpURLConnection) MDS_HOST_URL.openConnection();
  
  MDS_CONN.setRequestProperty("Content-Type", "application/xml");
  MDS_CONN.setRequestMethod("POST");
  
  MDS_CONN.setAllowUserInteraction(false);
  MDS_CONN.setDoInput(true);
  MDS_CONN.setDoOutput(true);
  
  String Message = "";
  Message += "\r\n";
  Message += "http://www.wapforum.org/DTD/pap_2.0.dtd\"\r\n []>\r\n";
  Message += "\r\n";
  Message += "\r\n";
  Message += "
  \r\n";
  Message += "\r\n";
  Message += "\r\n
  

• Question about REPLUSER (Session Variable)

  Hi all
  
  I try to produce a custom report to view all opportunities / activities generated by the respective user login ID, below is the detail & filter that I use in my report
  Subject: Opportunity Hostory / activity history
  Filter: Username (owned by the user) is equal to or in NQ_SESSION. REPLUSER
  
  I have 8 occasions and 4 activities created by a user in Siebel On Demand, but when I apply the filter above and I have connection user-a illustrates report No. result in retrurn. I missed something? Please advice
  
  Thanks in advance
  FINA

  Hi Fina!

  You must use the VALUEOF() function to get the value of NQ_SESSION. REPLUSER.
  For example, VALUEOF (NQ_SESSION. REPLUSER) will give you the login of the user who runs the report.

  Then in your filter, you must use the used. "" Employee Login " column (which in french translates as the e-mail address of the user...).
  It will look like used. Employee Login"= VALUEOF (NQ_SESSION. REPLUSER).

  I did a test and it worked for me, it shows the activities that I own.

  I hope this will help, do not hesitate to ask for more!

  Max

• Killed the remaining sessions in the session $ v

  Hello
  
  There are 42 sessions that are KILLED but still present in the view v$ session. v$ Resource_limit must keep 300 sessions. My question is: are these sessions killed count for example if session $ v shows 250 lines and 50 are marked Killed. 200 sessions will be counted or 250?
  
  Secondly, how can I KILL all open or multiple sessions or find them at the level of the Unix operating system which process id to kill at the level of the OS also?
  

  select sid,serial#,osuser from v$session where status='KILLED';
  
         SID    SERIAL# OSUSER                                                       
  ---------- ---------- ---------------
       56      8542 SYSTEM     
  - - 
  42 rows selected

  Thank you very much in advance.
  
  Concerning

  Is it possible that I have kill all sessions that are marked as being KILLED to CHANGE the SYSTEM COMMAND?

  No such statement available, you have to kill one by one with alter system kill session command but you can generate all the cmd like

  select 'alter system kill session ' || '''' || sid || ',' || serial# || '''' || ' immediate;' from v$session
  where status='KILLED';
  

  There is a PROCEDURE that is created for the normal user to kill user sessions

  put Word IMMEDIATE in alter system kill command inside the procedure.

• Analytic Workspace session not getting killed

  Hi all
  
  I created and mapped to a time dimension in the analytic workspace the user and proceeded to fill the option "keep the Dimension". For some reason, it took longer than expected (more than 15 minutes to the usual 5 minutes) to fill the size so I killed the session using Enterprise Manager ("marked to kill Session").
  
  Unfortunately, it's been more than three hours and the session is always contained in the EM performing an activity (Status: KILLED).
  
  "The AWM client is unable to fix another session resulting in error ' java.sql.SQLException: ORA-33290: analytic workspace < string > cannot be fixed in the mode that you asked because another user has attached in an incompatible mode.
  
  Questioning shows view v$ aw_olap an existing session into "READ WRITE" attach with changing mode permanently the values for the columns to display.
  
  Even to truncate the underlying dimension table does not work even if there is no locked in the database objects.
  
  How can I fix it? Any help will be appreciated.
  
  FYI my setup is as follows:
  11.1.0.7.0 DB (SUN Solaris 10)
  11.1.0.7.0 client (Windows Server 2008)

  Kill - 9

• session is not killed properly

  I use oracle 10g (10.2.0.1) in Redhat Ent Linux 5.0.

  In my production database, one of my session which is connected thorugh 'TOAD' program is hanged.

  I kill this session several times but not removed from the session and it shows the status is 'DEAD '.

  Here is the result:

  SQL > select username sid, serial #, session state $ v where status = 'KILLED ';

  USERNAME SID SERIAL # STATUS

  ------------------------------ ---------- ---------- --------

  1860 37381 KILLED DBPROD

  SQL > alter system kill session '1860,37381';

  alter system kill session ' 1860,37381'

  *

  ERROR on line 1:

  ORA-00031: session marked for kill

  SQL > alter system kill session '1860,37381' immediate;

  alter system kill session '1860,37381' immediate

  *

  ERROR on line 1:

  ORA-00031: session marked for kill

  Also during the session to kill, there is no show message in the log file of the alerts associated with the kill command.

  Now, please help me how I can fix this without closing the database.

  EDIT the SYSTEM KILL SESSION does not remove the session.  The session dies only when the user actually tries another SQL operation and is thrown a mistake - that is to say when Oracle kills the process.

  The other option is to use an OS command to kill the server process - but you need to know how to do this and identify the right process to kill (assuming not shared server and dedicated server connections).

  Hemant K Collette

• delete the entry for killed / sniped v$ session meeting

  Hi all
  In our database (11.2.0.1.0) of production running on windows server 2008 several client session list inactive last hour of 4-5 and I want to kill this session because that all sessions are dedicated to do this so I've created a set of resource plan his idle time after certain interval session status shows killed , but he is not removing of session entries $ v but our customer demand is there must be clear v $ session so for this I used command BACK utility taskkill to kill the particular operating system process, so I want to know what is the smart way to do it.

  We're talking about behavior, not a problem.

  In a dedicated server environment when the session is killed and the rollback ends the process so a join between v$ session and v$ process will fail. The v$ entered session also disappears usually fairly quickly, although I've seen cases where the v$ session entrance hung around until the instance has rebounded. But normally the Oracle will overlap the v$ session entry with a new session using the same sid but another serial # seconds on a busy system.

  Mark @ http://dbaspot.com/oracle-server/40419-killed-sessions.html

  Doc-id 1023442.6 MetaLink is also something confirm it.

  Also see link below and the last reply by Mr. Braj Kishore Mahto.
  http://dbaforums.org/Oracle/index.php?showtopic=3039

  Concerning
  Girish Sharma

  Published by: Girish Sharma on November 21, 2012 17:35

  So what is preferable in this respect:

  ALTER SYSTEM DISCONNECT SESSION

  The ALTER SYSTEM DISCONNECT SESSION syntax is an alternative method for killing Oracle sessions. Unlike the KILL SESSION command requesting the session to kill himself, the DISCONNECT SESSION command kill the dedicated server process (or virtual circuit when the shared use of Sever), equivalent to kill the operating system server process. The basic syntax is similar to the KILL command SESSION with the addition of the POST_TRANSACTION clause. The SID values and serial of the meeting concerned can be replaced in one of the following instructions.

  SQL > ALTER SYSTEM DISCONNECT SESSION ' sid, serial # ' POST_TRANSACTION;
  SQL > ALTER SYSTEM DISCONNECT SESSION ' sid, serial # ' IMMEDIATE;

  http://www.Oracle-base.com/articles/Misc/killing-Oracle-sessions.php#disconnect_session

• alter system kill session

  Hello
  
  While killing a session, I get a message like
  
  QL > alter system kill session '802,30515' immediate;
  alter system kill session '802,30515' immediate
  *
  ERROR on line 1:
  ORA-00031: session marked for kill
  
  I checked after a few hours again, I get the sid and the n serial form v$ session
  
  
  but the status of the user of the user is as 'dead '.
  
  SQL > select user name, session status $ v where sid = 802;
  
  STATUS OF USER NAME
  ------------------------------ --------
  APPS KILLED
  
  
  If the session was killed or not?
  
  Thank you
  Vijay

  887400 wrote:
  Its Linux X 86 64-bit

  But for some sessions, they will be killed immediately after alter system kill session command.,.

  but in this case it is not end immediately?

  What could be the cause?

  Internal Oracle session remains as long as the remains of operating system processes.

  Kill - 9

  Since you have already issued ALTER SYSTEM KILL SESSION
  You can't get the process ID of the Oracle's OS.

  You can associate up to the session Oracle with OS time as evidenced by time of CONNECTION
  ps - ef | LOCAL grep

• the call firefox.exe - ProfileManager does NOT open the Profile Manager does NOT create a new profile, KILLS the old session tabs, but all pinned!

  FF 7.0.1 on Win7 pro 64

  A behavior more bug / odd:
  calling "firefox.exe - ProfileManager".
  -do NOT open the Profile Manager,
  -does NOT create a new profile,
  -seems to do NOTHING (no change around records profile etc.).
  but...
  -KILLS tabs former session.
  -KILLS old session and leaves nothing to restore,
  -KILL all my pinned tabs!

  OK, I should downgrade to FF4.0.1, or same 3.6.x?
  These releases are much more reliable than 6 + 7 running!

  Try using this:
  
  Firefox.exe Pei