oracle11i 100% full even delete file
 

I am having problem with my system that make me confuse. When I run df -h I get
/dev/sda1 483G 455G 3.1G 100% /oracle11i

But when I do du -sh oracleiii I get another answer
[root@ebs3 ~]# cd /oracle11i
[root@ebs3 oracle11i]# du -sh
124G .

I did reboot of the server after deleting some files but the problem remain.

Where is going my space and how can I recover the space?

Please help me to fix this problem because I my system is down.

Regards

macuacua

Most probably, some process(es) (from Oracle, in all likelihood) opened large files, then removed their directory entries. You end up with "hidden" files - files with no names that continue to occupy space until the process closes them or exits. This is a common practice for temporary files: They are cleaned up automatically when the process that created them disappears.

df uses a different mechanism to calculate storage usage than du. df checks the filesystem's used space statistics, whereas du goes down a filesystem hierarchy, finds all files and sums up their sizes. Since du doesn't find the "hidden" files mentioned above, it reports less space used.

Such files have a "link count" (that is the number of names of a file) of 0. You can find them this way:

NLINK is the link count.

You probably recover your space when you stop the database.

Hi,

I did 2 times restart. I did database rebbot and the problem was not solve. Later, I did shutdown of database and reboot the server, even that the problem remain.
It is why I think I am looking for another procedure to fix problem.

may be it can help you to see what can be done-

Hi,

Hi,

Someone have answer for this issue?

macuacua

Can you put [code][/code] tags around your output? It is hard to read without any formatting.

Mr
 

Hi Ben,

I upload one file

Mr
 

Hi,

I upload in text file. If is not in readable, please advise.

Thank

Macuacua

I see a 2GB file, which is large, but nothing that explains your "loss" of about 300GB. So, right now I wonder what that might be.

Can you start the server without Oracle and check again? Also run a file system check on /dev/sda1.

What type of filesystem is that? I would like to see the superblock; perhaps it contains clues.

By the way, in this forum you can edit what you wrote (there is an Edit button), then add [code] before the output, and [/code] after it, so that it is formatted in a readable way.

Hi Beni,

In fact it is strange issue that I have.
The disks is come from netapp storage and I checked for an snapshot or something else, nothing.
Yesterday I did reboot server and poweroff (remove power cable) and I will repeat this procedure after one hour.

Did you identify which PID I can kill to get back those 2G?
On that filesystem there are oracle database.

Thanks

Nothing strange about this at all. Oracle reserves disk space for things, and rebooting/deleting won't necessarily help. What does your DBA say about this? Simple way to check things is to bring up your server with Oracle shut down, and see what happens to your disk space.

And have you contacted Oracle?? Oracle 11i is a commercial, pay-for product, and you get support when you pay. They can assist you with this. Also, as said, use CODE tags when posting things (don't put pictures or upload huge files), and don't bump your own thread after a short time, asking for help. If you're in a hurry, use the paid-for support from Oracle.

Hi,

What can I did shotdown now is bad

SQL> conn /as sysdba
ERROR:
ORA-09925: Unable to create audit trail file
Linux Error: 30: Read-only file system
Additional information: 9925
ORA-09925: Unable to create audit trail file
Linux Error: 30: Read-only file system
Additional information: 9925

URGENT

This looks like you somehow mounted your filesystem read-only. What does the mount command say?

DISCLAIMER: I know a little bit about Linux, but close to nothing about Oracle.

You can first read the "How to ask a smart question" link...this is NOT URGENT for anyone here. We volunteer our time, so asking for/expecting 'urgent' help is plain rude. Secondly, you were given pieces of advice:

• Restart the system WITHOUT ORACLE RUNNING, to see what happens to your disk space
• Ask your DBA for assistance
• CALL ORACLE SUPPORT

Have you done any of these things?? Your Oracle DB is down, because it is either misconfigured, running on a partition that's too small, or both. AGAIN: get your DBA to look at how Oracle is configured, and what it's doing. Call Oracle support, because **AGAIN** Oracle is not free. Oracle won't run when it's out of disk space, anymore than you can log in to your Linux system when the disk is 100% full.

Oracle's knowledgebase has many articles about this. Have you run disk cleanup?
https://blog.rackspace.com/just-the-...cle-disk-space