Out of Memory issue (was: redhat-list Digest, Vol 47, Issue 15)

[Imed Chihi]

On Thu, 2008-01-17 at 12:00 -0500, redhat-list-request at redhat.com wrote:
> Message: 17
> Date: Thu, 17 Jan 2008 14:12:31 +0000
> From: Andrew.Bridgeman at corusgroup.com
> Subject: Out of Memory issue
> To: redhat-list at redhat.com
> Message-ID:
> 	<OF8755B39E.89DA5FE6-ON802573D3.004D75A4-802573D3.004DEA5A at corusgroup.com>
> 	
> Content-Type: text/plain; charset=ISO-8859-1
>
> I currently have a problem with our Redhat Machines running out of memory
> and automatically killing the main process that is running on the machine
> at the time. Below is my kernel version and the text from the messages
> file. I have checked the top command and the machine seems to have loads of
> the 2 gig memory left. Could someone please let me know how to resolve this
> issue.
>
> Kernel - 2.4.21-20.ELsmp
> 
> Jan 17 09:21:13 kernel: Mem-info:
> Jan 17 09:21:13 kernel: Zone:DMA freepages: 2908 min:   0 low:   0
> high:   0
> Jan 17 09:21:13 kernel: Zone:Normal freepages: 1275 min: 1278 low: 4543
> high: 6303
> Jan 17 09:21:13 kernel: Zone:HighMem freepages:  249 min:  255 low: 4606
> high: 6909
> Jan 17 09:21:13 kernel: Free pages:    4432 (  249 HighMem)
> Jan 17 09:21:13 kernel: ( Active: 503137/73, inactive_laundry: 0,
> inactive_clean: 0, free: 4432 )
> Jan 17 09:21:13 kernel:  aa:0 ac:0 id:0 il:0 ic:0 fr:2908
> Jan 17 09:21:13 kernel:  aa:208678 ac:2066 id:73 il:0 ic:0 fr:1275
> Jan 17 09:21:13 kernel:  aa:289520 ac:2873 id:0 il:0 ic:0 fr:249
> Jan 17 09:21:13 kernel: 0*4kB 4*8kB 5*16kB 2*32kB 3*64kB 0*128kB 2*256kB
> 1*512kB 0*1024kB 1*2048kB 2*4096kB = 11632kB)
> Jan 17 09:21:13 kernel: 1*4kB 1*8kB 0*16kB 1*32kB 1*64kB 1*128kB 1*256kB
> 1*512kB 0*1024kB 0*2048kB 1*4096kB = 5100kB)
> Jan 17 09:21:13 kernel: 1*4kB 0*8kB 0*16kB 1*32kB 1*64kB 1*128kB 1*256kB
> 1*512kB 0*1024kB 0*2048kB 0*4096kB = 996kB)
> Jan 17 09:21:13 kernel: Swap cache: add 0, delete 0, find 0/0, race 0+0
> Jan 17 09:21:13 kernel: 2061 pages of slabcache
> Jan 17 09:21:13 kernel: 166 pages of kernel stacks
> Jan 17 09:21:13 kernel: 0 lowmem pagetables, 1634 highmem pagetables
> Jan 17 09:21:13 kernel: Free swap:      0kB
> Jan 17 09:21:13 kernel: 524160 pages of RAM
> Jan 17 09:21:13 kernel: 294768 pages of HIGHMEM
> Jan 17 09:21:13 kernel: 10422 reserved pages
> Jan 17 09:21:13 kernel: 10491 pages shared
> Jan 17 09:21:13 kernel: 0 pages swap cached
> Jan 17 09:21:13 kernel: Out of Memory: Killed process 22262
> (mpp970_s_5434a_).

In this instance at least you do seem to have ran out of memory for
real: you got only 5 MB of lowmem (~0.5%), 1 MB of highmem and no free
swap at all.

The lower_zone_protection parameter does not exist in RHEL 3, so you
cannot use it there.

The hugemem kernel will not help you in instances like the above, it
will only make your system a bit slower. The 64-bit mode won't help
either.

One thing you can consider doing is adjust the overcommit parameters so
the VM does not make too many allocation "promises" which it cannot
keep.

 -Imed

-- 
Imed Chihi - عماد الشيحي
Senior Technical Account Manager         Office: +44-1-252-362-962
http://www.redhat.com/promo/vendor     Mobile: +44-7-921-700-936