Re: [cobalt-users] RaQ3i Memory Leak?

[Mat Kovach <mkovach@xxxxxxxxxxxxxxxxx>]

On Sun, Mar 05, 2000 at 02:11:33AM -0500, Craig N. wrote:
: LOL... I'm catching up Debbie.. It's been only 4-6 hours since I rebooted,
: and I'm back down to:
: 
:              total       used       free     shared    buffers     cached
: Mem:        128180      66464      61716      89832       8980      29088

Why are you rebooting, just because the memory is not marked as free? Is the
system halting or not responding?

He is an older system I used from my personal site, but some of a few
friends.  It is just a 486dx2/66 with 16MB.

mek:/root # free -t
             total       used       free     shared    buffers     cached
Mem:         14472      13280       1192      17460        524       3568
-/+ buffers/cache:       9188       5284
Swap:        33772       7764      26008
Total:       48244      21044      27200  

Note that while the system has "used" all the physical memory, it has barely 
touched the swap and has been running for over almost 30 days (previous uptime 
was 200+, had to move the machine).  Also the load is 1+ because I run the 
distributed.net client on it.

The important part is to understand HOW Linux uses memory and what free
is actually reporting on. 

When disk space is used in Linux, that just means it has been allocated 
for a purpse: run programs, shared memory, buffers, or cachen.  It does not
mean it is unavailable.

Where you run into problems, is when you system is consuming more and more
slower swap (which is has limited speed based on disk I/O).

: down 20000 in just 4 hours.. <grin>  I could understand it if this machine
: actually had more than 15 domains/users, but these domains/pages aren't even
: being hit.. and it just seems to leak more and more K as the hours pass,
: with or without connections being made to the box...  I'm sure I'll catch
: you at 6040k by this time tomorrow if I don't reboot again.. LOL..  I had
: another RaQ3i (that was replaced) and it had the same problem with memory..
: I'm just not sure if it's indicative to the RaQ3 (or Linux/Apache in
: general) or possibly PHP3, and/or SSH (mySQL isn't running, I put the devil
: to bed until he's needed).

No, you don't understand how memory management in Linux works.  Was you system
checking on sway space, completely unresponsive, giving kernel panics? 
I'd suggestion getting a decent book on Linux and learning a bit on how
it handles memory.

-- 
Mat Kovach                                      mkovach@xxxxxxxxxxxxxxxxx
Cleveland Linux User Group                       http://cleveland.lug.net