maximum o/s file size?

[Robert Nowak_2]

I am a DBA that is unloading a large database table to a flat file that appears to be hitting some kind of O/S limitation. I unload the table as root user, who has ulimit set to "unlimited", when it stops once it hits 2GB. We are running HP-UX 11.

[Vincenzo Restuccia]

#fsadm -o largefiles /mount_point

[A. Clay Stephenson]

Hi,

You need to enable 'largefiles'. I assume this is a vxfs filesystem so
fsadm -F vxfs -o largefiles /fsmountpoint

if HFS substitute -F hfs

[Bill McNAMARA_1]

I believe not only do you need to create the fs or convert it to tthe largefiles option but you also need to mount it with the same option.

Later,
Bill

[Robert Nowak_2]

What are the consequences to running this command below:

fsadm -F vxfs -o largefiles /fsmountpoint

Will it erase/destroy any files on that filesystem that I'm converting from nolargefiles to largefiles? Is this a command that takes effect immediately? Also, will it retain this characteristic the next time the machine is rebooted?

Is there any benefit to you when I assign points to your answer? I'm new to this forum as of today.

[Patrick Wallek · Accepted solution]

There will be no effect on files currently in the filesystem when you run the fsadm command. You can do it on the fly and it will take effect immediately. To make sure the large files option is retained after a reboot, check your /etc/fstab and make sure that the LV is mounted with the largefiles option.

As to the points issue - The benefit to us is that it adds to our totals, but the main reasoning behind the points system is that it lets future users know which answer helped the most in solving your problem.

[A. Clay Stephenson]

The only change is that files > 2GB are allowed. On vxfs filesystems, the command can be run while the filesystem is mounted and the change is immediate. Largefiles will be allowed until a fsadm -F vxfs -o nolargefiles /fsmountpt command is issued and is preserved over system reboots.

The only consequence is if there are files over 2GB on a filesystem and when the filesystem is mounted and the -o nolargefiles mount option (not fsadm) then the mount will fail. The -o nolargefiles mount option must be removed.

On an hfs filesystem, the filesystem must not be mounted and you use the filesystem block device rather than the mountpoint.

Hope this clears it up, Clay