Is there any way this limit could be raised for sparc64? I use sparc64
quite extensively both personally and professionally and it would be
nice to have this limit raised for the entire arch without having to
maintain custom kernels on many different boxes. Is there any reason to
not raise this limit to at least match amd64? If there is security or
performance implications in making a change then I understand, but if it
is simply an arbitrary setting, it would be nice to have more suitable
defaults. In addition, I am also regularly bumping up against the amd64
limit, and have to go out of my way to unnecessarily parallel-ize tasks,
or thrash the disk just for the sake of working around the issue. I
can't be the only person who would like to make a ram disk bigger than
32GB (or 8GB on sparc64).
With fully loaded modern T-series machines becoming available on ebay at
price points attainable to mere mortals, I can see this limit becoming
more of an issue. It is somewhat disheartening to have to use sort -T on
a file set and fill my disk with temp files while still having over a
dozen times the MAXDSIZ available in unoccupied memory.
On 01/30/18 22:35, Ted Unangst wrote:
Jordan Geoghegan wrote:
Is there any particular reason for the low sparc64 MAXDSIZ? Is there any
way for this limit to be increased as I have some large data
manipulation that needs to be done and I really would love to be able to
increasing the value can be done by changing the header and compiling a new
kernel. larger values should work, but apparently nobody has needed them until
now.
