Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations strongm on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

how to 'really' clear out dump area?

Status
Not open for further replies.

aking

Technical User
Aug 11, 2002
112
GB
Hi.
I'm running Oracle 10gR2 on windows 2003
Could someone explain what exactly does the message "78% of dump area used" mean in Enterprise Manager? - and how to make it go away??

i have cleared out the bdump directory - it had about 83MB, i left the alert log (obviously) and the latest traces, now that directory is down to 6.5MB (the alert log is 5.5MB).
I have also cleared out the udump directory - even tho it was only 6MB.
The cdump directory is empty.
In the path d:\oracle\product\10.2.0\admin\<dbasename>\ i also have directories called "pfile" - 3KB, "dplog" - 1KB, and "backup" - 6GB. I have not touched those directories as i think they have nothing to do with the dump area, is this correct?

if i do 'show parameter max_dump_file_size' it says 'unlimited'.

So does anyone know what this message means? 78% of what is used? Why hasn't Enterprise Manager noticed that i have reduced the files in bdump and udump? I deleted these files 2 days ago - when the alert was on 71%, yet it is going up each day.
Oh yeah, i have about 63GB free space on this drive. The flash area max size is 40GB, currently 16GB are used - so i have enough space.
The database backs itself up every night - complete backup as it's quite small - these have been successful. Also i run a schema export every night - also has been successful. There are currently no archivelogs visisble anywhere - rman deletes them after successful backups. So the only thing visible in the flash area are three days worth of backups - this is all exaclty as i would expect it to be.

i hope this is enough background. It is probably obvious from this that i'm not an oracle expert - and i'm not a fan either. But i do need to keep this database going and i'm worried it's going to stop when it gets to 100% because of some arcane oracle reason. We have a deadline of 17 december, if the database stops before then i may was well write my cv! Any help very much appreciated, this is my 2nd day searching - otherwise i wouldn't be posting such questions as i'm sure this is simple.
 
This is the space taken up by your whole database - try to set up some housekeeping tasks to keep it stable.

The internet - allowing those who don't know what they're talking about to have their say.
 

Also clean-up "audit" directory, network logs and trace files ($TNS_ADMIN/logs, $TNS_ADMIN/trace).
[3eyes]


----------------------------------------------------------------------------
The person who says it can't be done should not interrupt the person doing it. -- Chinese proverb
 
Thanks anyway but these answers don't mean anything to me.
The space taken up by my database (according to enterprise manager is) 6.3GB.
I have a tiny database and tons of space (60GB free).
Any out of disk space errors are gonna be down to Oracle thinking it has much less room than it really does. That's what i want to rectify.
I know there are housekeeping tasks which run every day. I know my database is stable - it's been up pretty consistently for 3 years.
The 'audit' directory is empty, (d:\oracle\product\10.2.0\<dbasename>\RDBMS\audit). I've looked for all directories called 'trace' they are all empty. I've looked for all directories called 'log' or 'logs' - also empty. I thought clearing out the udump directory was clearing out the traces.
But most of all i'm not worried about how many traces or logs or whatever else i've got. I'm looking at things the other way around. I have GB's of space that i'll never need. I want to tell Oracle to go ahead and use whatever space it needs so that it doesn't stop.

 

From SQL*Plus do "show parameter dump" and examine all the folders listed.
[tongue]

----------------------------------------------------------------------------
The person who says it can't be done should not interrupt the person doing it. -- Chinese proverb
 

PS: Click on the 78% and it will show which areas it is referring to.
[wink]

----------------------------------------------------------------------------
The person who says it can't be done should not interrupt the person doing it. -- Chinese proverb
 
Thanks guys, but i'm pretty sure that my dump area isn't full.
'show parameter dump' give the three default locations for bdump, cdump and udump.
I know all these areas and i've emptied them out - total space taken in these directories is 6MB as stated in first post.

i'm hoping it is either because i'm on windows and oracle doesn't know how much space i've got and is confused or a bug with enterprise manager - which i know can be out of date with its alerts.
Either way i want to get rid of an alert which is increasing daily - especially before it gets to 100% - as i'm wondering if Oracle would continue to write to the alert log if that happened....

 

If you are on WinDoze, to release space "Empty Recycle Bin" will do the trick.
[medal]

----------------------------------------------------------------------------
The person who says it can't be done should not interrupt the person doing it. -- Chinese proverb
 
i've got upwards of 30GB free - i'm pretty sure this is not a space issue.
i've got 2 servers, on one of them the bin was already empty, i emptied it on the other one for consistency, overnight the % of dump area used has increased by 1%.
I think it is something to do with how alerts work on Oracle. I have logged a call with metalink - for 4 days now so far. It's the most difficult support for a product i've ever used. I can't even get an answer to direct questions - like can Oracle accurately assess free space on windows, and if so how does it do it.
Right now i wonder if it can.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top