Export Problem - Hanging

Hi all,

Oracle Ver : 8.1.7.2
OP : Sun Solaris 2.7

Problem:-

I've tried exporting a table which has 23 million records (3.1gb) twice and both times it hung at the same point - I put feedback on and it stops at 15 million records (587mb - compressed). I've create a pipe and compressing the pipe to my dump file....

Also, when I try and do a select Count(0) from this table the whole db runs very slow, nearly to a halt...

I've tried the dbverify command on each of my datafiles and it reported no corruption... When not selecting on this table the db is fine with no problems at all


Any ideas????

:confused:

I the gzip actually transfering from the pipe to the .dmp . or is the pipe just growing and growing and eventually filling the filesystem?

When running the exp, make sure the .dmp file in with you are piping to is growing and not only the pipe...

Specifically:

nohup gzip < gzip_pipe > exportfile.dmp.gz &

Cheers...

the pipe is working properly - I was watching the pipe grow then back to 0 and the dmp file was getting bigger and bigger....

Could the table be corrupt?? - Select count(0) is killing the db!

How do I check to see if it is?

Thanks in advance

analyze table table_name validate structure;

As my db is in CHOOSE mode, but runs in RULE. If I analyze a table it will force it into COST and the db will run like a dog?!

Is there a way around this??

Quote:

---

Originally posted by krw
Hi all,

Oracle Ver : 8.1.7.2
OP : Sun Solaris 2.7

Problem:-

I've tried exporting a table which has 23 million records (3.1gb) twice and both times it hung at the same point - I put feedback on and it stops at 15 million records (587mb - compressed). I've create a pipe and compressing the pipe to my dump file....

Also, when I try and do a select Count(0) from this table the whole db runs very slow, nearly to a halt...

I've tried the dbverify command on each of my datafiles and it reported no corruption... When not selecting on this table the db is fine with no problems at all


Any ideas????

:confused:

---

Amrgon - Any advice?

It won't actully calculate the statistics.

But for future reference you can...

analyze table table_name delete statistics;

Well, let me ask the obvious.

Are you blowing out the mount point?

MH

No - we are only using 37%

Is there anything information in the alert.log? Or a .trc in the udump directory?

Nothing in alert file or .trc files...

I'm currently analyzing table.. Let you know!!

grjohnson - Where abouts in york are u? I'm also in york!!

Quote:

---

Originally posted by krw
Nothing in alert file or .trc files...

I'm currently analyzing table.. Let you know!!

grjohnson - Where abouts in york are u? I'm also in york!!

---

PM me and I'll let you know.

Quote:

---

Originally posted by krw
grjohnson - Where abouts in york are u? I'm also in york!!

---

OK, but NO touching!

Interested in the company, thats all!!

Anyway, on a work note :- I tried to analyze the table but after 1 hour it was still running but wasn't getting anywhere!! It wasn't active anymore!!

Any ideas with what I can do?!

Thanks in advance...

Can you do a

select * from table_name where rownum < 50;

It may have a corruption in the table... but a block corruption is usually reported during a FTS or an analyze. Have you got indexes on the table?

This is strange :-

After trying to do a select count(0) with rownum < 50 (which worked) I kept upping the value until I did a FTS with select count(0) and IT WORKED!!!

Could the analyze of fixed this do you think?! I'll try another export and see what happens...?

Has the instance been restarted since yesterday?

Anyway, let us know how you get on.

Cheers,

Greg.