[rhelv6-beta-list] My first experiences with RHEL6 beta

[John Summerfield]

Bryan J. Smith wrote:
> John Summerfield wrote:
>> When I do want swap, I usually want a swap file, not a
>> swap partition.  Swap files are more flexible, they can
>> be created and dispensed with at will.  Its a minor task
>> to add a swap file, provided only that disk space is
>> available. It's even easier to dispense with swap and
>> assign the recovered space to other tasks.
> 
> Most of this argument is mitigated thanx to volume management
> (i.e., LVM).
> 
> As far as swap file v. swap partition, that goes into the same
> category of filesystem free reservations, segmentation of
> filesystems, etc...  You are free to prefer what you wish, but
> "best common practices" of 40 years of POSIX/POSIX-like
> implementations tend to favor some sound reasons why we have
> such things as:  
> - swap partitions
> - filesystem free reserevations
> - segmentation of filesystems
> - etc...
> 
> And, again, liberal usage of LVM tends to make most arguments
> on the matter quite moot in "Enterprise" Linux.  My $0.02.
> 

I would venture to suggest that most systems running Linux, especially 
those with IA32 or AMD-64 CPUs, have a single disk. I cannot see that 
LVM provides any benefit in such cases.

I used to use OS/2. If you check googlism.com you will find vestiges of 
my reputation there. on OS/2, which never had swap partitions, the 
recommend placement for swap is "the busiest partition on the least busy 
drive."

The "least busy drive" needs no explanation, but people do tend to choke 
on their weeties at "the busiest partition."

reflect a moment. At a random instant, where are the drive's heads 
likely to be? I suggest over the latest read or write operation. Where 
is that, usually? Someplace in the busiest partition.

It does not matter what OS you use, the above is true.

Now, with default partitioning on any Linux distro I have seen, where 
the user chooses "one partition for everything," there are two partitions.

One for the data, covering almost all the drive.

One for the swap, either on the inside edge or the outside edge of the 
disk, it makes little difference.


Now, I do not understand how Linux filesystems decide where on the disk 
to create new files. Likely it either starts from one perimeter and 
works across the drive (for simplicity, let's assume new disks like when 
you've just formatted them, reusing space complicates matters) like 
Windows does, or it allocates them in band like OS/2 with HPFS does.

Maybe it depends on which filesystem is used.

If it uses the first technique, there's an even chance that one's 
initial files are as far as possible from the swap partition. Unless the 
folk thought of this and designed their installers appropriately, but I 
am not ready to assume they have.

If the technique is more like the second, then the situation is better, 
but still any access to data ensures the heads are removed from the swap 
partition by quite an amount, and any access to swap ensures the heads 
are quite a distance from the next data to be read or written.

Enterprise systems, of course, are different. Enterprises can choose to 
buy and install a Very Expensive Disk to support hi I/O traffic.

btdt. I worked at the Australian Dept of Social Security when it 
implemented the original Medibank in the 1970s. We bought a Very 
Expensive Computer system (IBM's finest at the time), and and Especially 
Expensive Disk to hand hi I/O traffic.

Not SWAP/Paging though, the highest I/O load was the index for the 
database. Surprised me;-)

But I'm not talking about enterprise systems.


-- 

Cheers
John

-- spambait
1aaaaaaa at coco.merseine.nu  Z1aaaaaaa at coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375

You cannot reply off-list:-)