Re: monitoring listener.log
Title: monitoring listener.log hey can u elaborate how do u scan each log file, and convert it in email msg. how do u do this.. saurabh sharma dba - Original Message - From: NGUYEN Philippe (Cetelem) To: Multiple recipients of list ORACLE-L Sent: Wednesday, April 25, 2001 9:10 PM Subject: monitoring listener.log Hi list, We have a script wich monitor the alert.log for each database. This script scan the alert.log file and send an email if an ORA-xxx error occured Now, we want to write a script monitoring the listener.log file : wich kind of message or variable indicate a network problem (ORA-xxx, ...) ? ...then we 'll have the same question for the log file made by OEM ! TIA
Re: monitoring listener.log
Hi, Typical errors in sql*net traces are like this: nserror: nsres: id=0, op=67, ns=12571, ns2=12560; nt[0]=533, nt[1]=57, nt[2]=0 Means: You get protocol error 57. If you are using TCP/IP, check winsock errors, if I'm not wrong it should be 20057 or 10057. This error causes sql*net 533 error. 533,12560 and 12571 are result. Saurabh Sharma wrote: hey can u elaborate how do u scan each log file, and convert it in email msg.how do u do this.. saurabh sharmadba - Original Message - From: NGUYEN Philippe (Cetelem) To: Multiple recipients of list ORACLE-L Sent: Wednesday, April 25, 2001 9:10 PM Subject: monitoring listener.log Hi list, We have a script wich monitor the alert.log for each database. This script scan the alert.log file and send an email if an ORA-xxx error occured Now, we want to write a script monitoring the listener.log file : wich kind of message or variable indicate a network problem (ORA-xxx, ...) ? ...then we 'll have the same question for the log file made by OEM ! TIA
Re: monitoring listener.log
Hi again, Note that listener is only used while creating shadow processes on server side. Once shadow processes are created by listener, listener is never used. in other words, listener process is not a bridge between user process and server process during data transmissions. So, scanning listener.log only is not a good way to catch sql*net errors. to catch sql*net errors after connections are established, set sql*net server trace. then scan them. But, it's not be efficient. there will be a lot of trace files. rgrds... "Danisment Gazi Unal (Unal Bilisim)" wrote: Hi, Typical errors in sql*net traces are like this: nserror: nsres: id=0, op=67, ns=12571, ns2=12560; nt[0]=533, nt[1]=57, nt[2]=0 Means: You get protocol error 57. If you are using TCP/IP, check winsock errors, if I'm not wrong it should be 20057 or 10057. This error causes sql*net 533 error. 533,12560 and 12571 are result. Saurabh Sharma wrote: hey can u elaborate how do u scan each log file, and convert it in email msg.how do u do this.. saurabh sharmadba - Original Message - From:NGUYEN Philippe (Cetelem) To: Multiple recipients of list ORACLE-L Sent: Wednesday, April 25, 2001 9:10 PM Subject: monitoring listener.log Hi list, We have a script wich monitor the alert.log for each database. This script scan the alert.log file and send an email if an ORA-xxx error occured Now, we want to write a script monitoring the listener.log file : wich kind of message or variable indicate a network problem (ORA-xxx, ...) ? ...then we 'll have the same question for the log file made by OEM ! TIA
RE: monitoring listener.log
Title: monitoring listener.log Hi and thanks to " Danisment Gazi Unal (what does it mean ?)" Here is the script : #!/bin/sh # This script make 2 things : # 1- archive alert files # 2- search "ORA-" in alert.log file # if error is found , send an email to adresses passed in parameters # # Entry parameters : # 1- SID # 2- list of emails separated by a space # Ex : % ALTERTCHECK ORCL [EMAIL PROTECTED] [EMAIL PROTECTED] # --- SCRIPT_DIR=/home/ORCL/script ORACLE_SID=$1 shift EMAIL=$* # Set ALERT_DIR and ALERT_FILE # -- ALERT_DIR=/app/oracle/admin/$ORACLE_SID/bdump ALERT_FILE=alert_$ORACLE_SID.log ALERT_LOG=$ALERT_DIR/$ALERT_FILE ERRFILE=$SCRIPT_DIR/ALERT_ERRORS.TXT SUBJECT="Error in alert file of $ORACLE_SID" # Remove the alert file #30 # if [ -f $ALERT_LOG.30 ] then /bin/rm $ALERT_LOG.30 fi # Renanme the alert file, ex: 29 to 30, 28 to 29, .., 01 to 02, etc. # -- i=29 while [ $i -ge 1 ] ; do j=`expr $i + 1` if [ -f $ALERT_LOG.$i ] then mv $ALERT_LOG.$i $ALERT_LOG.$j fi i=`expr $i - 1` done # Rename the current alert log file to #1 # - if [ -f $ALERT_LOG ] then mv $ALERT_LOG $ALERT_LOG.1 fi # Create a new alert file, not mandotory (Oracle will do this) # touch $ALERT_LOG chmod 777 $ALERT_LOG # Remove the error file # --- if [ -f $ERRFILE ] then /bin/rm $ERRFILE fi # Searching for ORA- in the alert log #1 # if error, send mail to adresses in parameters # --- cat $ALERT_LOG.1 |grep '\ORA-' > $ERRFILE if [ -s $ERRFILE ] then cat $ERRFILE | mailx -s "$SUBJECT" $EMAIL fi -Message d'origine-De: Saurabh Sharma [mailto:[EMAIL PROTECTED]]Date: vendredi 27 avril 2001 12:12À: Multiple recipients of list ORACLE-LObjet: Re: monitoring listener.log hey can u elaborate how do u scan each log file, and convert it in email msg. how do u do this.. saurabh sharma dba - Original Message - From: NGUYEN Philippe (Cetelem) To: Multiple recipients of list ORACLE-L Sent: Wednesday, April 25, 2001 9:10 PM Subject: monitoring listener.log Hi list, We have a script wich monitor the alert.log for each database. This script scan the alert.log file and send an email if an ORA-xxx error occured Now, we want to write a script monitoring the listener.log file : wich kind of message or variable indicate a network problem (ORA-xxx, ...) ? ...then we 'll have the same question for the log file made by OEM ! TIA
