On Thu, 2003-06-05 at 17:26, Gordon Messmer wrote:
> Skip Morrow wrote:
> > Quick and dirty: Why don't the memory usages reported for each program
> > from 'ps -aux' and/or 'top' (even after pressing the "H") add up to the
> > totals as being reported by 'top' and/or 'free'?  Try it and you'll see
> > what I mean.
> 
> Variety of reasons:
> 
> Some processes share memory.  Different instances of the same program 
> will share code pages.  Different programs entirely will share code 
> pages when they use the same libraries.  Multi-threaded programs will 
> share both code and data pages.  This sharing makes the sum of the 
> memory sizes too big.
> 
> The X server will mmap your video card's memory, so its size appears 
> considerably bigger than it really is.
> 
> You may also be looking at the wrong piece of output from 'free' or 
> 'top', entirely.  The "total" memory use reported by 'free' includes 
> application memory use as well as disk buffers and cache.
> 
> 

Well, here's the problem (I should probably start a new thread for this,
because it may turn out to be very eye-opening)

I had noticed that I didn't have much free memory a few days ago (I had
384M RAM installed) so I went and bought another 256 and installed it
(totalling 640M)  Restarted the computer and saw that I was only using
20% of the RAM.  But after a few hours, I noticed that the little gnome
bar graph was getting pretty high, so I rechecked and it was up to 85%. 
And a few hours later, it was over 95% and I had started using swap
space.  So I thought "memory leak".  I started shutting down everything
I could, rebooted a lot, and still find that no matter what I do, the
computer slowly eats up more and more RAM (10-20 megs per hour).  I even
turned off every service, rebooted (less than 30 total processes
running) and still had the RAM slowly getting used up.  I tried it with
kernel 2.4.20-18.9 (the newest, and the other kernels that have been
released since RH9.0 came out.  As it is now, I am rebooting once or
twice a day to keep the RAM usage down, which is mildly inconvenient :(

Anyone have any ideas here?

Skip

Here's the SHIFT-SCRLK output

**Fresh boot**
[EMAIL PROTECTED]:~$ Mem-info:
Free pages:      596032kB (     0kB HighMem)
Zone:DMA freepages: 13072kB min:  4224kB low:  4352kB high:  4480kB
Zone:Normal freepages:582960kB min:  2296kB low: 11252kB high: 16240kB
Zone:HighMem freepages:     0kB min:     0kB low:     0kB high:     0kB
Free pages:      596032kB (     0kB HighMem)
( Active: 8291/33, inactive_laundry: 0, inactive_clean: 115, free:
149008 
)
4*4kB 6*8kB 5*16kB 4*32kB 4*64kB 4*128kB 1*256kB 1*512kB 1*1024kB
5*2048kB 
= 13072kB)
0*4kB 4*8kB 1*16kB 0*32kB 0*64kB 0*128kB 1*256kB 0*512kB 1*1024kB 
284*2048kB = 582960kB)
= 0kB)
Swap cache: add 0, delete 0, find 0/0, race 0+0
Free swap:       459004kB
163836 pages of RAM
0 pages of HIGHMEM
3067 reserved pages
15687 pages shared
0 pages swap cached


**after a few hours of running**
[EMAIL PROTECTED]:~$ Mem-info:
Free pages:       61380kB (     0kB HighMem)
Zone:DMA freepages:  5404kB min:  4224kB low:  4352kB high:  4480kB
Zone:Normal freepages: 55976kB min:  2296kB low: 11252kB high: 16240kB
Zone:HighMem freepages:     0kB min:     0kB low:     0kB high:     0kB
Free pages:       61380kB (     0kB HighMem)
( Active: 83425/2, inactive_laundry: 18223, inactive_clean: 2483, free: 
15345 )
93*4kB 37*8kB 14*16kB 7*32kB 1*64kB 1*128kB 0*256kB 0*512kB 0*1024kB 
2*2048kB =
5404kB)
2392*4kB 1185*8kB 430*16kB 115*32kB 40*64kB 14*128kB 14*256kB 4*512kB 
2*1024kB 7*2048kB = 55976kB)
= 0kB)
Swap cache: add 1010, delete 131, find 639/640, race 0+0
Free swap:       455476kB
163836 pages of RAM
0 pages of HIGHMEM
3067 reserved pages
73375 pages shared
879 pages swap cached




-- 
redhat-list mailing list
unsubscribe mailto:[EMAIL PROTECTED]
https://www.redhat.com/mailman/listinfo/redhat-list

Reply via email to