file: table is full

[Camel_1]

The server just had the subject problem and user could not logged into the server. Finally we had performed a system dump (TC) and rebooted the server, it is only a database server running Oracle instances only, the nfile value is 7816.

I have performed the Q4 analyze of the dump, do you know how to check which process(es) used such large amount of open files?

Thanks in advance.

Simon

[Robert-Jan Goossens]

Hi simon,

Now you have rebooted it will almost be inpossible to find the culprate. Normaly you could check with sar -v or ps -ef | grep ora | wc -l if the system or oracle is hitting the nfile limit.

nfile at +/- 7500 is not very high. Below is good thread, check the Pete's and Bill's posts.

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=327287

Kind regards and a good weekend,

Robert-Jan

[Robert-Jan Goossens]

Hi simon,

Now you have rebooted it will almost be impossible to find the culprate. Normaly you could check with sar -v or ps -ef | grep ora | wc -l if the system or oracle is hitting the nfile limit.

nfile at +/- 7500 is not very high. Below is good thread, check the Pete's and Bill's posts.

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=327287

Kind regards and a good weekend,

Robert-Jan

[Todd McDaniel_1]

Simon,

Depending on your amount of memory, you can raise nfile very high...

Mine is at 100k at the moment on my main Production box. Here is my sar -v output, even on a DEV box...

Showing nproc=16020, ninode=18648, & nfile=2959


# sar -v 1 10

HP-UX hostname B.11.11 U 9000/800 04/02/04

15:27:46 text-sz ov ..proc-sz ......ov ....inod-sz ......ov ....file-sz ov
15:27:47 N/A N/A ....905/16020 0 .....5995/18468 0 .....16395/29539 0

[Marvin Strong]

Are you tracking your systems performance? Patrol, GPM, etc...

You could maybe get a report from that historical data.