Re: Memory, swap, and limits

2008-06-21 Thread Bruno Wolff III
On Wed, Jun 18, 2008 at 09:50:37 -0700, Rick Stevens <[EMAIL PROTECTED]> wrote: > > If your RAM is heavily fragmented or heavily used, the system may find > it difficult to locate adequate contiguous RAM and spend a lot of time > swapping things to disk and back as tasks compete for the free RAM.

Re: Memory, swap, and limits

2008-06-19 Thread Gene Heskett
On Thursday 19 June 2008, tom wrote: >Sadly, many of us are so set in our ways and fictions >that we are pretty much asleep at school. Sadly, I find that is a trap I have tripped over more than once of late, which can be embarrassing in front of all these frogs. :) -- Cheers, Gene "There are f

Re: Memory, swap, and limits

2008-06-19 Thread tom
On Thu, 19 Jun 2008, Gene Heskett wrote: On Thursday 19 June 2008, James Kosin wrote: [...] Sorry, I'm a bit dated and now I've probably revealed my age. I've been around since the first personal computers and remember when 64K was considered a LOT of memory. James So was I James, but heck,

Re: Memory, swap, and limits

2008-06-19 Thread Bruno Wolff III
On Wed, Jun 18, 2008 at 15:58:49 +, Beartooth Sciurivore <[EMAIL PROTECTED]> wrote: > > OK; but then there's something else I don't know, or don't > understand. Does a Fedora machine do any swapping while it has memory > left? I didn't think I had (or needed) any control at all over

Re: Memory, swap, and limits [now becoming OT]

2008-06-19 Thread fred smith
On Thu, Jun 19, 2008 at 08:57:18AM -0400, James Kosin wrote: > Bruno Wolff III wrote: > >On Wed, Jun 18, 2008 at 09:56:25 -0400, > > James Kosin <[EMAIL PROTECTED]> wrote: > > > >>> > >>> > >>Having TOO much swap space can be a detriment and not an asset. > >>Usually, the rule of thum

Re: Memory, swap, and limits

2008-06-19 Thread Gene Heskett
On Thursday 19 June 2008, James Kosin wrote: [...] >Sorry, I'm a bit dated and now I've probably revealed my age. >I've been around since the first personal computers and remember when >64K was considered a LOT of memory. > >James So was I James, but heck, I remember Pearl Harbor, about when the t

Re: Memory, swap, and limits

2008-06-19 Thread James Kosin
Bruno Wolff III wrote: On Wed, Jun 18, 2008 at 09:56:25 -0400, James Kosin <[EMAIL PROTECTED]> wrote: Having TOO much swap space can be a detriment and not an asset. Usually, the rule of thumb I go by is allocate about 2x the amount of physical memory installed on the system;

Re: Memory, swap, and limits

2008-06-19 Thread Bruno Wolff III
On Wed, Jun 18, 2008 at 09:56:25 -0400, James Kosin <[EMAIL PROTECTED]> wrote: >> > Having TOO much swap space can be a detriment and not an asset. > Usually, the rule of thumb I go by is allocate about 2x the amount of > physical memory installed on the system; for machines with < 1M. T

Re: Memory, swap, and limits

2008-06-18 Thread Patrick O'Callaghan
On Wed, 2008-06-18 at 09:50 -0700, Rick Stevens wrote: > The portion of a process' code that is to be executed must be in RAM, > along with any data structures it may need (unless they're the > un-mmap(2)d parts of files). If there is inadequate contiguous space > in RAM, idle processes will be sw

Re: Memory, swap, and limits

2008-06-18 Thread Patrick O'Callaghan
On Wed, 2008-06-18 at 11:06 -0500, Mikkel L. Ellertson wrote: > > OK; but then there's something else I don't know, or don't > > understand. Does a Fedora machine do any swapping while it has > memory > > left? I didn't think I had (or needed) any control at all over > swapping, > > beyond

Re: Memory, swap, and limits

2008-06-18 Thread Rick Stevens
Beartooth Sciurivore wrote: On Wed, 18 Jun 2008 08:58:47 -0500, Mikkel L. Ellertson wrote: Beartooth Sciurivore wrote: [...] Otoh, I've never gotten anywhere near filling up a hard drive, except once when I had a testbed machine triple booting three different distros. So why c

Re: Memory, swap, and limits

2008-06-18 Thread Mikkel L. Ellertson
Beartooth Sciurivore wrote: OK; but then there's something else I don't know, or don't understand. Does a Fedora machine do any swapping while it has memory left? I didn't think I had (or needed) any control at all over swapping, beyond choosing how much space to afford it. What I have see

Re: Memory, swap, and limits

2008-06-18 Thread Beartooth Sciurivore
On Wed, 18 Jun 2008 08:58:47 -0500, Mikkel L. Ellertson wrote: > Beartooth Sciurivore wrote: [...] >> Otoh, I've never gotten anywhere near filling up a hard drive, >> except once when I had a testbed machine triple booting three different >> distros. So why can't I at least increase

Re: Memory, swap, and limits

2008-06-18 Thread James Kosin
Mikkel L. Ellertson wrote: James Kosin wrote: Having TOO much swap space can be a detriment and not an asset. Usually, the rule of thumb I go by is allocate about 2x the amount of physical memory installed on the system; for machines with < 1M. This number will need to approach more or less

Re: Memory, swap, and limits

2008-06-18 Thread Mikkel L. Ellertson
James Kosin wrote: Having TOO much swap space can be a detriment and not an asset. Usually, the rule of thumb I go by is allocate about 2x the amount of physical memory installed on the system; for machines with < 1M. This number will need to approach more or less 1x for machines with 1-2M.

Re: Memory, swap, and limits

2008-06-18 Thread Mikkel L. Ellertson
Beartooth Sciurivore wrote: Every new computer I've yet had has begun slowing down soon after I get it -- probably because I keep several browsers open, with from several to many tabs each. I've learned long since to make sure each machine has all the memory it can handle from the git-go, befo

Re: Memory, swap, and limits

2008-06-18 Thread James Kosin
Beartooth Sciurivore wrote: Every new computer I've yet had has begun slowing down soon after I get it -- probably because I keep several browsers open, with from several to many tabs each. I've learned long since to make sure each machine has all the memory it can handle from the git-go, befo

Memory, swap, and limits

2008-06-18 Thread Beartooth Sciurivore
Every new computer I've yet had has begun slowing down soon after I get it -- probably because I keep several browsers open, with from several to many tabs each. I've learned long since to make sure each machine has all the memory it can handle from the git-go, before it ever reaches m