Isn't it true that the closer to the edge of the
disk, the faster the filesystem there will be
accessed? It's not that big of an improvement,
but every little bit helps, especially in the case
of swapping processes in and out of memory
on a heavily loaded server. swap is not a file-
system, but the physical location in proximity to
the edge of the disk still applies (hence the default
locations of / and swap, s0 and s1 respectively).
I personally would rather have a slice rather than
a swapfile, if at all possible. You still can't make
a swap partition larger than 2gb on a single slice,
right? Or is there a patch to fix this?
>> Recently, with 9gig+ disks becoming more and more prevalent,
>> when I install Solaris 2.6 or 7, I'm just creating a root
>> file system (ie, /dev/dsk/c0t0d0s0), and a file system for
>> the swap space (ie, /dev/dsk/c0t0d0s1).
>> Since these machines frequently suffer the "scorched earth" policy,
>> I don't have to worry about users filling up "/var" or "/opt", etc.,
>> before it's time for a reinstall. Much nicer.
>Just out of curiosity, what's a scroched earth policy?
>> I wondering, though, what the performance ramifications would
>> be if I went one step further, and didn't even create a "swap"
>> filesystem, but just used the "mkfile" command, to create
>> a 1 gig, or so, swapfile after installation, adding it to
>> /etc/vfstab?
>Bare in mind that crash dumps get saved on the primary swap device.
>If it's a swap file on a filesystem that can't be mounted, you'll
>be knackered!
>Personally, I'd stick with your current partitioning scheme,
>although there's nothing wrong with having an additional swap
>file on another disk (having another swap file on the same disk
>makes no sense from a performance point of view).
>HTH,
>--
>Rich Teer
>NT tries to do almost everything UNIX does, but fails - miserably.
>The use of Windoze cripples the mind; its use should, therefore, be
>regarded as a criminal offence. (With apologies to Edsger W. Dijkstra)
>If it ain't analogue, it ain't music.
>Voice: +1 (250) 763-6205
>WWW: www.rite-group.com