Experimental Physics and
Industrial Control System

"J. Lewis Muir" <[email protected]> · Tue, 14 Oct 2008 16:08:43 -0400

On 10/14/08 3:30 PM, Maren Purves wrote:
> Tom Cumming wrote:
>> I've hit the 2GB Channel Achiver index file limit. How have people been
>> dealing with this limit? Is it time to move it up to 4GB? Modern
>> machines should be able to handle it.
> 
> I don't think this has so much to do with the channel archiver as with a
> 2 GB file size (addressing) limit on 32 bit machines. We have a similar
> problem with (non-EPICS related) data files (in our case it means
> limiting the exposure times on non-destructive readout observations
> (one read per second, all go into the same file)).

I don't know about file size limits possibly imposed by Channel
Archiver, but I think 32-bit machines can support files larger than 2GB;
it depends on the OS and the file system.

For example, with MS Windows 2000 or XP, even FAT32 can support files as
large as (2^32)-1 bytes (1 byte less than 4 GB).
Ref: <http://support.microsoft.com/default.aspx?scid=kb;EN-US;314463>

As another example, with RHEL 3 (based on the Linux 2.4 kernel), ext3 is
supposed to support files as large as 2 TB.
Ref: <http://www.redhat.com/rhel/compare/>

-lewis

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System