Tim Smith wrote:
> In article <1218062.m9IdtGTYQt@xxxxxxxxxxxxxxx>,
> Roy Schestowitz <newsgroups@xxxxxxxxxxxxxxx> wrote:
>> > There again goes Hadron Quarks claim of "easy to allocate a
>> > non-fragmented swap file". Which would be quite a bit larger than
>> > 256meg
>>
>> What else would you expect from Tim "filesystem expert", "weekly stats
>> analyst", "Funkenbush protege" Smith?
>
> It doesn't take a filesystem expert to figure out that you were wrong
> about Linux filesystems. It just takes using the "filefrag" command.
>
> Or it just takes paying attention. Peter has pointed out many times
> that fragmentation is a problem when you try to create large files.
No, fragmentation is not a "problem", since linux does not suffer from it
It *is* a problem when creating swap files
> This is why, if you are going to use a swap file, you should create it
> soon after the filesystem is created, if you want it to be contiguous.
Even then, it /can/ be difficult, as the fs has already had several
write/read/delete cycles during install
I have always maintained that swapfiles are fine to add *temporarily* swap
space. They still have disadvantages to swap partitions.
You can't "Suspend to RAM" in swap files for example (although maybe this
has changed meanwhile. Linux evolves much faster than windows)
> If you are going to advocate Linux, maybe you should put a little effort
> into learning more about it?
>
Heed your own advice, Tim Hadron
--
Tact, n.:
The unsaid part of what you're thinking.
|
|
