Re: What slices should I allocate on 2nd drive during installation?

In article <Iu47LF.F2p@xxxxxxxxxxxxxxxx>,
Charles Lindsey <chl@xxxxxxxxxxxxxxxx> wrote:
In <drud88$kn9@xxxxxxxxxxxxxxxxxxxxxxx> jgp@xxxxxxxxxxxxxxxxxxxxxxx (Jim Prescott) writes:
My current practice is to put a 30M (meg, not gig) s7 partition on
all disks just for use as a LVM/DiskSuite metadb area. If you ever
decide you want to setup RAID or mirroring, you'll be glad it's there.
You also need to put some of your swap space on the 2nd disc. My
understanding is that, however much swap space you decide to have, ideally
it should be split evenly around all the discs available.

He may want to put swap there but "need" is too strong. Ideally all
disk traffic is split evenly among all the disks available; swapping is
only a part of that.

One example, say you know a particular application is going to be
largely diskbound on a particular disk. Putting swap on that disk too
will just slow down the application, and might even slow down all
swapping since now the system will be doing part of its swapping onto a
very busy disk.

Trying to speed up swap is also usually a minor optimization. Ideally
you have enough RAM so swapping has only minor performance impact. If
your system's performance is heavily influenced by swap speed you will
almost always see bigger payoffs from trying to reduce the swapping
rather than speed it up. Sometimes it just isn't possible to have
enough RAM so striping swap will help some, but this is likely to only
turn a very slow system into a slow system.

But in the OPs case where the second disk appears to be mainly archival
space (and thus presumably rarely accessed), spreading swap onto the
disk would probably be a win.
--
Jim Prescott - Computing and Networking Group jgp@xxxxxxxxxxxxxxxxxx
School of Engineering and Applied Sciences, University of Rochester, NY
.

Relevant Pages

  • Re: OOM killer gripe (was Re: What still uses the block layer?)
    ... not thrashing is that all of the pages in RAM are in the swap cache, ... used to be able to do 3 point something megabytes per second to/from disk, ... But swapping heavily has been a viable mode of operation ... speeds aren't keeping up with processor and memory increases. ...
    (Linux-Kernel)
  • Re: High memory pressure makes system nearly unresponsive.
    ... Disk io has priority. ... Slackware, the swap usage grows to about 600MB, and UI interactivity ... I don't think it could have much to do with swapping, ...
    (comp.os.linux.misc)
  • Re: ~500 megs cached yet 2.6.5 goes into swap hell
    ... Total number of kilobytes the system paged in from disk per second. ... >> Swapping out is good, but when that's coupled with swapping in as is the ... it creates a thrashing situation where we swap out to ... the system swaps in and out more while cache ...
    (Linux-Kernel)
  • Re: Page fault, GEOM problem??
    ... > filesystems so the swap has to be enabled. ... > As to why you can write kernel core dumps only to certain devices the ... > check is performed whether the driver for the disk you want to dump on ...
    (freebsd-stable)
  • Re: dynamic allocation of swap disk space
    ... Increasing swap can allow more physical RAM to be used for ... > caching data from disk. ...
    (Linux-Kernel)