Re: Space under $ORACLE_BASE/$ORACLE_HOME keeps increasing why
From: onedbguru <onedbguru_at_yahoo.com>
Date: Sat, 19 Mar 2011 15:00:10 -0700 (PDT)
Message-ID: <9e07633d-fc6b-484c-a8ff-32677e6c7e7e_at_dn9g2000vbb.googlegroups.com>
On Mar 19, 10:50 am, zigzagdna <zigzag..._at_yahoo.com> wrote:
> I am on Oracle 10.2.0.4 on HP UNIX 11.2. I am running a program which
> has an infinite loop. Each loop iteration sleeps for 15 seconds, does
> some datbase access using jdbc…? I am noticing that space under
> $ORCALE_BASE keeps increasing. My program is definitely not writing
> any files under $ORACLE_BASE, so Oracle is creating some files, else
> how can I explain increase in $ORACLE_BASE. When I stop the program,
> space shrinks back to original, so looks like file has been deleted.
>
> My program could have some cursors not being closed, but this will
> increase the memory of Oracle shadow process and not space under
> $ORACLE_BASE. Are there any memory mapped files, where shared memory
> is being mapped to some file under $ORACLE_BASE. I am unable to find
> any such files using:
> Find . –mtime -1
>
> I see some Oracle files such as Oracle agent log, listener.log but do
> not see anything unusual. Can someone explain what is going on and how
> can I troubleshoot this problem.
>
> I am on Oracle 10.2.0.4 on HP UNIX 11.2. I am running a program which
> has an infinite loop. Each loop iteration sleeps for 15 seconds, does
> some datbase access using jdbc…? I am noticing that space under
> $ORCALE_BASE keeps increasing. My program is definitely not writing
> any files under $ORACLE_BASE, so Oracle is creating some files, else
> how can I explain increase in $ORACLE_BASE. When I stop the program,
> space shrinks back to original, so looks like file has been deleted.
>
> My program could have some cursors not being closed, but this will
> increase the memory of Oracle shadow process and not space under
> $ORACLE_BASE. Are there any memory mapped files, where shared memory
> is being mapped to some file under $ORACLE_BASE. I am unable to find
> any such files using:
> Find . –mtime -1
>
> I see some Oracle files such as Oracle agent log, listener.log but do
> not see anything unusual. Can someone explain what is going on and how
> can I troubleshoot this problem.
>
> java [program does various selects, inserts, updates; it writes its
> execution log under /gps/twq/db3 and
> ORACLE_BASE is /gpsg/twq/oracle
>
> bdf shows...
>
> /dev/vg03/lvtwqstage
> 8192000 6088123 1972403 76% /gpsg/twq/stage
> /dev/vg03/lvtwqoracle
> 12288000 8418465 3637466 70% /gpsg/twq/oracle
> /dev/vg03/lvtwqdb1 8192000 4752239 3225929 60% /gpsg/twq/db1
> /dev/vg03/lvtwqdb4 8192000 5435984 2583829 68% /gpsg/twq/db4
> /dev/vg03/lvtwqexpbkup
> 72466432 22210280 47115159 32% /gpsg/twq/expbkup
> /dev/vg04/lvtwqdb2 66322432 26020448 37783114 41% /gpsg/twq/db2
> /dev/vg04/lvtwqdb3 66322432 26936468 36924350 42% /gpsg/twq/db3
> /dev/vg06/sq_db1 65536000 50530190 14067949 78% /gpsg/getsq/db1
>
> It is /gpsg/twq/oracle which keeps increasing until program is
> recycled, then frees up space under /gpsg/twq/oracle..
>
> I ran
> lsof /gpsg/twq/oracle | sort -k 7n,7
>
> It shows an Oracle shadow process 1778 size is 1519m and this size
> keeps increasing. What I do not undesrtand why
> this shadow process is writing to this directory and what eaxctly it
> is writing which occupies so much disk sapce.
>
> COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
> oracle 29334 twqora txt REG 64,0x30002 262845312 3633 /gpsg/twq/oracle/
> product/10.2.0/bin/oracle
> oracle 1778 twqora 7u REG 64,0x30002 1519238927 28811 /gpsg/twq/oracle
> (/dev/vg03/lvtwqoracle)
Date: Sat, 19 Mar 2011 15:00:10 -0700 (PDT)
Message-ID: <9e07633d-fc6b-484c-a8ff-32677e6c7e7e_at_dn9g2000vbb.googlegroups.com>
On Mar 19, 10:50 am, zigzagdna <zigzag..._at_yahoo.com> wrote:
> I am on Oracle 10.2.0.4 on HP UNIX 11.2. I am running a program which
> has an infinite loop. Each loop iteration sleeps for 15 seconds, does
> some datbase access using jdbc…? I am noticing that space under
> $ORCALE_BASE keeps increasing. My program is definitely not writing
> any files under $ORACLE_BASE, so Oracle is creating some files, else
> how can I explain increase in $ORACLE_BASE. When I stop the program,
> space shrinks back to original, so looks like file has been deleted.
>
> My program could have some cursors not being closed, but this will
> increase the memory of Oracle shadow process and not space under
> $ORACLE_BASE. Are there any memory mapped files, where shared memory
> is being mapped to some file under $ORACLE_BASE. I am unable to find
> any such files using:
> Find . –mtime -1
>
> I see some Oracle files such as Oracle agent log, listener.log but do
> not see anything unusual. Can someone explain what is going on and how
> can I troubleshoot this problem.
>
> I am on Oracle 10.2.0.4 on HP UNIX 11.2. I am running a program which
> has an infinite loop. Each loop iteration sleeps for 15 seconds, does
> some datbase access using jdbc…? I am noticing that space under
> $ORCALE_BASE keeps increasing. My program is definitely not writing
> any files under $ORACLE_BASE, so Oracle is creating some files, else
> how can I explain increase in $ORACLE_BASE. When I stop the program,
> space shrinks back to original, so looks like file has been deleted.
>
> My program could have some cursors not being closed, but this will
> increase the memory of Oracle shadow process and not space under
> $ORACLE_BASE. Are there any memory mapped files, where shared memory
> is being mapped to some file under $ORACLE_BASE. I am unable to find
> any such files using:
> Find . –mtime -1
>
> I see some Oracle files such as Oracle agent log, listener.log but do
> not see anything unusual. Can someone explain what is going on and how
> can I troubleshoot this problem.
>
> java [program does various selects, inserts, updates; it writes its
> execution log under /gps/twq/db3 and
> ORACLE_BASE is /gpsg/twq/oracle
>
> bdf shows...
>
> /dev/vg03/lvtwqstage
> 8192000 6088123 1972403 76% /gpsg/twq/stage
> /dev/vg03/lvtwqoracle
> 12288000 8418465 3637466 70% /gpsg/twq/oracle
> /dev/vg03/lvtwqdb1 8192000 4752239 3225929 60% /gpsg/twq/db1
> /dev/vg03/lvtwqdb4 8192000 5435984 2583829 68% /gpsg/twq/db4
> /dev/vg03/lvtwqexpbkup
> 72466432 22210280 47115159 32% /gpsg/twq/expbkup
> /dev/vg04/lvtwqdb2 66322432 26020448 37783114 41% /gpsg/twq/db2
> /dev/vg04/lvtwqdb3 66322432 26936468 36924350 42% /gpsg/twq/db3
> /dev/vg06/sq_db1 65536000 50530190 14067949 78% /gpsg/getsq/db1
>
> It is /gpsg/twq/oracle which keeps increasing until program is
> recycled, then frees up space under /gpsg/twq/oracle..
>
> I ran
> lsof /gpsg/twq/oracle | sort -k 7n,7
>
> It shows an Oracle shadow process 1778 size is 1519m and this size
> keeps increasing. What I do not undesrtand why
> this shadow process is writing to this directory and what eaxctly it
> is writing which occupies so much disk sapce.
>
> COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
> oracle 29334 twqora txt REG 64,0x30002 262845312 3633 /gpsg/twq/oracle/
> product/10.2.0/bin/oracle
> oracle 1778 twqora 7u REG 64,0x30002 1519238927 28811 /gpsg/twq/oracle
> (/dev/vg03/lvtwqoracle)
Look under $ORACLE_BASE/admin... audit (*.aud) and trace files (*.trc, *.trm) files, depending on definition of db_recovery_file_dest, audit_file_dest, background_dump_dest, alert log, listener logs and a host of other files will placed in the their respective "destinations" - generally defaulting somewhere under $ORACLE_BASE. Also, if your database has a lot (or any) triggers, the base 11.2.0.1 has a problem where the test team forgot to remove a (trace forever level 12?) command in the actual code causing Oracle to create a trace file EVERY time a trigger fires. There is a patch to remove this code. Search "metalink" Received on Sat Mar 19 2011 - 17:00:10 CDT