[OpenAFS-devel] 1.4.0 largefile support question

[Horst Birthelmer]

On Dec 6, 2005, at 10:59 AM, Peter Somogyi wrote:
> We were testing the --enable-largefile-fileserver option for 1.4.0.
> We have found this feature unusable for both SLES9 and AIX 5.2, too  
> - in 1.4.0.
> I'm writting here because we need urgent answer about the desired  
> availability of this option, and our customer is waiting for the  
> result.
>
> For AIX, we've found one (and further more) trivial bug for "vos  
> dump", see http://rt.central.org/rt/Ticket/Display.html?id=23607.
>
> For linux, we have found that writting a file of 6 GB fails (at the  
> position of 4GB):
>> dd if=/dev/zero of=6gb_file bs=1MB count=6144
>>
>> File size limit exceeded
>> stoneblade11:/afs/.afscell/largemount # ls -sh
>> total 4.1G
>> 2.0K .  2.0K ..  4.0G 6gb_file
>

 From the context above, I'm not sure, what you're trying to do.
You're calling 'dd' in an AFS directory and the volume mounted there  
is on a AIX fileserver, right?

You're sure the client has a 'largefile enabled' AFS?
You're sure the server can write more that 4GB?
Try doing that directly on the device on the server. (/viceXX) Just  
the same 'dd' call for making sure your server is able to write more  
than 4GB.
On AIX you have to enable the large file support on JFS as well in  
order to have the fileserver writing large files.

If the large file support of AFS would have failed, your writing  
process usually gets stuck at 2GB.
.. and besides, I wrote more than 4GB into AFS on my machines before.

Horst