Ryan,
you may also want to read Sarah's blog about maxcontig
from back in time when she fixed/changed this:
http://blogs.sun.com/roller/page/sarahsblog?entry=inquiring_minds_want_to_know
---
frankB
On Thu, 25 Aug 2005 23:35:57 +0200, Sarah Jelinek <[EMAIL PROTECTED]>
wrote:
Hi Ryan,
yes, maxcontig can be any positive integer value. maxcontig as of
Solaris 10 and Solaris 9 update 7, was decoupled from the i/o transfer
size so that we could make it as large as the user wanted.
So, when you say, 'allocating single sequential block extents', do you
mean ensuring the reservation of this sequential set of blocks as well?
thanks,
sarah
*****
Matty wrote:
Howdy,
I am doing some testing, and wondered if the value assigned to
maxcontig can be larger than the number of logical blocks in a cylinder
group? The mkfs_ufs man page would lead me to believe the answer to my
question is "yes":
maxcontig=n The maximum number of logical
blocks, belonging to one file, that
are allocated contiguously. The
default is calculated as follows:
[ .... ]
I want to see if it's possible to configure UFS to simulate VxFS's
ability to allocate single sequential block extents. We have several
applications that perform large sequential I/O's, and have
significantly increased throughput by allocating one large sequential
extent per file. Can we achieve the same thing by adjusting cgsize and
maxcontig with UFS? Any thoughts or feedback are welcome!
Thanks,
- Ryan
_______________________________________________
ufs-discuss mailing list
[email protected]
_______________________________________________
ufs-discuss mailing list
[email protected]
_______________________________________________
ufs-discuss mailing list
[email protected]
