frustrating memory leaks

bird603568 · 08-04-2005, 08:03 AM

I found it its kded it killed it and gained a lot back.

xushi · 08-05-2005, 06:43 AM

This may sound weird, but ....

How hardware damaging is turninig off your swap while still in X and there's ~190M swap usage?

My HD acted as if it failed on me today.. the usual CLINK CLINK CLINK noise, hardly any response from it.. even while booting linux. After leaving it off for a bit and back on, and its back to normal again (although i dont trust it.. i've had me a fair ammount of HD crashes and i know one when i see one)

Got a Hitachi HD checking utility and it says its fine.. which is the major headfsck.. heh. Anyway (drifted there for a bit).. Could my swap issue be from a bad HD ? And could this abormal HD behavior be caused from my swapoff /dev/hda2 ?

xushi · 08-10-2005, 05:02 AM

I've switched to slamd64 to see if it was something with my current OS.. same problem.

I've switched to gentoo to see if it was anything slackware related, same problem..

I haven't compiled anything with gtk2 because i was told gtk2 can cause mem leaks, stil... same problem

I've replaced my 256 stick with a 1Gig stick of ram, same problem...

Now my RAM usage goes up to 490Mb after an hour of usage, and just now my Swap is at 2.6Mb. (did i mention i have 1.25 gigs of ram?) If i keep this running till night time, i bet you the swap usage will go up to a few hundred megs.

i'm really upset from this, and i stil can't pinpoint the problem... anything i run... be it FF, gaim, xchat, amule... anything... if i keep it on and running for a few hours, it starts to leak. Here's the latest top if it helps.

Code:

top - 11:01:55 up  2:09,  3 users,  load average: 0.34, 1.12, 1.13
Tasks: 139 total,   2 running, 137 sleeping,   0 stopped,   0 zombie
Cpu(s):  6.3% us,  2.7% sy,  0.0% ni, 90.0% id,  0.7% wa,  0.3% hi,  0.0% si
Mem:   1282748k total,  1175996k used,   106752k free,   121296k buffers
Swap:   248996k total,     2684k used,   246312k free,   576044k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND                                
 8279 root      15   0  101m  36m  14m S  4.3  2.9   2:14.00 amule                                   
 8181 root      14  -1 83688  61m 8140 S  1.0  4.9   1:12.57 X                                       
 8314 root      15   0 1477m 197m  26m S  0.7 15.8   0:08.99 java                                    
 8333 root      15   0 1477m 197m  26m S  0.7 15.8   0:23.44 java                                    
 8335 root      15   0 1477m 197m  26m S  0.7 15.8   0:51.13 java                                    
 8331 root      15   0 1477m 197m  26m S  0.3 15.8   0:06.01 java                                    
 8418 root      15   0 1477m 197m  26m S  0.3 15.8   0:24.66 java                                    
 8433 root      15   0 1477m 197m  26m S  0.3 15.8   0:16.21 java                                    
 5871 root      15   0 47592  29m  13m S  0.3  2.4   0:10.79 firefox-bin                             
17203 root      16   0 10684 1324  952 R  0.3  0.1   0:00.02 top                                     
    1 root      16   0  2624  544  456 S  0.0  0.0   0:00.56 init                                    
    2 root      RT   0     0    0    0 S  0.0  0.0   0:00.00 migration/0                             
    3 root      34  19     0    0    0 S  0.0  0.0   0:00.18 ksoftirqd/0                             
    4 root      10  -5     0    0    0 S  0.0  0.0   0:00.10 events/0                                
    5 root      17  -5     0    0    0 S  0.0  0.0   0:00.00 khelper                                 
    6 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 kthread                                 
    9 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 kacpid                                  
   83 root      10  -5     0    0    0 S  0.0  0.0   0:00.00 kblockd/0                               
   86 root      16   0     0    0    0 S  0.0  0.0   0:00.00 khubd                                   
  153 root      15   0     0    0    0 S  0.0  0.0   0:00.00 pdflush                                 
  154 root      15   0     0    0    0 S  0.0  0.0   0:00.17 pdflush                                 
  156 root      20  -5     0    0    0 S  0.0  0.0   0:00.00 aio/0                                   
  155 root      15   0     0    0    0 S  0.0  0.0   0:00.81 kswapd0                                 
  745 root      15   0     0    0    0 S  0.0  0.0   0:00.02 kseriod                                 
  816 root      11  -5     0    0    0 S  0.0  0.0   0:00.00 ata/0                                   
  853 root      15   0     0    0    0 S  0.0  0.0   0:00.23 kjournald                               
  861 root      16   0     0    0    0 S  0.0  0.0   0:00.00 scsi_eh_0                               
  862 root      15   0     0    0    0 S  0.0  0.0   0:00.18 usb-storage                             
  916 root      12  -4  2624  540  416 S  0.0  0.0   0:00.07 udevd                                   
 5242 root      15   0     0    0    0 S  0.0  0.0   0:00.63 kjournald                               
 6680 root      15   0  4968  756  592 S  0.0  0.1   0:00.00 syslog-ng                               
 7272 rpc       16   0  9016  716  560 S  0.0  0.1   0:00.00 portmap                                 
 7316 root      11  -5     0    0    0 S  0.0  0.0   0:00.00 rpciod/0

piete · 08-10-2005, 06:13 AM

Swap will get filled up when the system runs out of RAM. You can think of it as an overflow carpark that's slightly further away from Tesco than the regular carpark =) Also, sometimes, people use the overflow carpark anyway. Swap is a useful way of moving memory resident things out of the fast memory and into the slower one - it may be something in your kernel (kernel = car park attendant, perhaps?) that is allowing swap usage before your actual memory is filled up. Beyond that, I'm just guessing.

So, let's examine some of this:

Quote:

Got a Hitachi HD checking utility and it says its fine

My understanding of checking utilities is that they simply check the disk sectors. If the HD has a more mechanical problem that is affecting the spin mechanism, then it's entirely possible that the utility wouldn't pick up on that kind of fault.

Quote:

Could my swap issue be from a bad HD ?

If you mean that in relation to "why are programs leaking into swap and then not going away when i close them?" then I seriously doubt it.

Quote:

And could this abormal HD behavior be caused from my swapoff /dev/hda2 ?

Again it seems unlikely that swapoff will cause such extreme hardware failure.

A few tests and a few queries:

* Boot into run level 3 and let it sit there for an hour without logging in - see what happens. If nothing much happens when you log in and check top, you can conclude your core OS is probably ok, and it's related to something other than the basic system.
* Do the same with run level 4. Just boot into it and leave it for an hour. Check mem usage after an hour. If nothing else happens, then X itself is not causing your issue.
* Now go back to run level 3 (always clean boot from power off, never reboot) and log in. Leave it for an hour, make sure it's nothing that's being run by users.
* Same with run level 4, log in and leave it for an hour.

Now you've eliminated the potential system problems - time to check the programs you're running, and how you're running them.

* amule uses java? I'm not familiar with amule, I'm afraid, but Java is notorious for not cleaning up after itself very well. Try using the PC *without* java for an hour and see what happens.
* I note that everything is running as root - try running things as a user instead (I can't imagine what kind of difference it will make, but it's one more test scenario that can be checked) - so into run level 3, login as a user, startx ... etc etc.

Good luck on getting everything back - let me know how it goes =)

- Piete.