Allocated swap space versus used swap space

Allocated swap space versus used swap space

Post by Marcelo Tosatt » Sat, 21 Apr 2001 15:10:07



Hi,

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in

More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

 
 
 

Allocated swap space versus used swap space

Post by Marcelo Tosatt » Sat, 21 Apr 2001 15:30:09



> Hi,

Argh. Silly.

Well...

Right now we can get a task killed by the OOM killer even if there is a
lot of _unused_ (but allocated) swap space. The reason for that is the
pre allocation of swap.

Practical example (128MB swap, 960MB ram):

# fillmem 960

...

   procs                      memory    swap          io     system cpu
 r  b  w   swpd   free   buff  cache  si  so    bi    bo   in    cs  us sy
 0  0  0      0 873548   1760   9244   0   0    25    11   32    24   0 1 98
 0  0  0      0 873540   1760   9244   0   0     0     0  103     8   0 0 100
 1  0  0      0 469040   1760   9248   0   0     2     0  221   562   2 17 80
 1  0  1  91980   2068     80  98480   0 362     0   456  277   724   2 26 72
 0  1  2 128516   1560     80 106564  30 24724    44 24724  443  8090   0 9  91
 0  1  0  64840 820804     84  66396 112 22436   148 22438  413  9197   1 13  87
 0  0  0  64840 820416    104  66768   0   0   198     0  119    22   0 0 100
 0  0  0  64840 820416    104  66768   0   0     0     0  102     9   0 0 100

...

Out of Memory: Killed process 763 (fillmem).

Around 60MB was written to swap, so there was unused swap space available,
which means the task should _not_ get killed.

We cannot simply check for "(swapper_space.nrpages > 0)" or count the
number of dirty swap cache pages to avoid the OOM killer from triggering,
obviously.

We can split the swap map in "used" / "allocated" bits (or something which
gives us that information), but that would be painful to do, I guess.

Comments ?

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in

More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

 
 
 

1. plenty of swap space, but keep receiving error "swap space limit exceeded"

Syslog reports
  unix: WARNING: /tmp: File system full, swap space limit exceeded

But

  Filesystem            kbytes    used   avail capacity  Mounted on
  swap                 2423720   48480 2375240     3%    /tmp

The host:
  SunOS <hostname> 5.7 Generic_106541-16 sun4u sparc SUNW,Ultra-60

I ran vmstat 1 for some time and watched for available swap space to
go to zero. But the lowest it would go is 75kB.

When analyzing this problem, I ran into the following error twice, but
for different commands:
  Can't run command <command>
  fork(2) failed; no more memory

What's going on?

Thanks.

2. Network problem.. RedHat 6.0 droping LAN connection>??>

3. Try to mount a seocn SWAP space-- error" unable to find swap space signature"

4. How do I configure NAT?

5. Is swap space in SCO only swap space or is it consider virtual memory

6. Reminder: This group is archived!

7. Adding swap space using 'swap -a' command ?

8. doom - 8-bit sound...HOW?

9. How much swap space should I allocate (Slackware) ?

10. Allocate Swap Space

11. Linux HardDrive Space Reqs, Swap Space

12. Paging space size (NOT swap space size)

13. /tmp Swap space running out of space