Help, Memory Leak Problem, Suggestions appreciated.

Help, Memory Leak Problem, Suggestions appreciated.

Post by Richard Lacroi » Wed, 23 Aug 1995 04:00:00

One of my users has a problem with memory leaks, we have an HP9000
running watchworks under HP-UX 9.05.
we suspect a memory leak, but don't know how to pursue it.

This machine has been up for a few days, and shows 96m of swap space
in use.  This figure seems to creep up indefinatly until rebooted.

Is there an effective way to see what processes have what memory,
and track it over time to see if a particular facility is
loosing memory?


Phone: (202)944-7306


1. HELP: Wierd memory leak problem

        Let me give some system info first.  My system has a 486SX-33, 8
megs RAM, 2 40 meg MFM HDs (don't laugh), trident SVGA card, mono VGA
monitor, floppy-chain QIC-80 tape drive, 1.44 floppy, SB & WSS clone sound
card, 2S1P1G, and a VLB IDE controller (currently only driving the floppy
chain).  I call it the "college student special" |->.
        I have installed $lackware Linux 2.1 (disk sets A, AB, and N) and
have cleaned up most of the configuration problems that the install left
me with.  I have set up an account for myself and gotten PPP to work
(yay!).  No wierd mods/kernels/whatever - aside from adding myself as a
user, installing IRCII, and changing some config files to make PPP work,
the system is virgin slackware (IDE kernel).  However, from the beginning
there has been an omnipresent problem.
        I'll boot up (from HD using MBR LILO), start doing stuff, and
after about 1/2 hour (it varies), the system will suddenly start accessing
the drive like mad and the system will slow to a virtual lockup (literal
MINUTES until a keypress is echoed).  The HD access will gradually die
down (but the system won't return to normal) for about 5 minutes, when the
HD accessing begins again and the system comes back to life for about 30
seconds and it starts all over again.
        I watched the memory usage on another console (all consoles slow
down, but switching between them doesn't) all the way from bootup to the
weirdness point on /proc/meminfo, and here's what I saw.  ~3.5 megs was
shown as used immediately after bootup and login as root.  No swap usage
yet (I allocated ~2.5 megs for the swap partition).  RAM usage rises
quickly to about 6.5 megs and then begins a slow climb upward at about
10K/minute or so, occasionally fluctuating up or down a bit.  Still no
swap usage.  When RAM usage hits about 40-50K below the ceiling, BAM!
Instantly, 1.5megs of swap is used - with NO reduction is RAM usage!
Needless to say, this is the point at which the 'troubles' begin.  I've
let the cycle described above run for ~30 minutes, and eventually things
just come to a halt - C-A-D won't even work, necessitating a cold reset
with all the attendant filesystem munging and repair on the next bootup.
        Any help would be appreciated.  I've read all the FAQs and HOWTOS
that apply to my system and have come up empty.  I'm also setting up the
same slackware distribution on a P90/32M RAM/2 540M HD/PCI/Mach64/CDROM
system at work, and I haven't noticed this problem there.  Replies by
E-mail are preferred.  Thanks in advance.


2. Linux - Windoze

3. memory leaks by shared memory and fork

4. Are you a FROG?

5. Sun workshop memory monitor does not detect all memory leak

6. Problems with Cisco router

7. 64bit solaris 2.8 memory name resolution memory leak?

8. What is the maximum partition size

9. How can I recover memory from a memory leak

10. mm/memory.c, 2.4.1 : memory leak with swap cache (updated)

11. How to find out the memory size used by a process or check its memory leaking?

12. Help please - strange FTP problem - any help much appreciated :)

13. Linux Workstation compatibility, suggestion appreciated ;-)