right now we keep having Oracle query to the MS SQL send from the Oracle gateway, some time it rans fast and sometimes it hold there for serveral days and make the tempdB full and MS SQL operate anymore:
I usually run the follow query to query the sys.sysprocess to see their login time and how long it has been ran. Then if I see a the login used by the oracle to query MS SQL and it takes much longer than expected, I will use the spid to kill that by:
SELECT spid, loginame, program_name, hostname as "From host", login_time, last_batch, DB_NAME(dbid) AS "Against database" ,
(SELECT text FROM sys.dm_exec_sql_text(sql_handle))as "Query executed"
FROM master..sysprocesses WHERE loginame= 'ixapi' and open_tran > 0
Any SP script you guys can introduce so that:
1) to build a SP that run that query by using the script above, and if the time it takes is more then one hour, it look at the SPID.
2) it create the kill commmand with the spid number as a string.
3) it execut the whole kill command with taht SPID we need one by one, if there are more than one result from the query about( it is rared to see but it did happen!)
4) email the statement, host, program name, loginname of each spid killed to the DBA (the operator is functional already) so that they know something anormal was happened.
any example is good for me to do it.