1. Home
  2. Cisco Systems
  3. Cisco 851/%SYS-2-MALLOCFAIL: Memory allocation of 65536 bytes failed

Cisco 851/%SYS-2-MALLOCFAIL: Memory allocation of 65536 bytes failed

Hi,

First time posting here.

I have a new Cisco 851 router. Like clockwork, every 4 days I find the following in the SYSLOG on my main server:

Dec 25 15:44:48 Cisco 000064: *Dec 25 09:54:01.476 PCTime: %SYS-2-MALLOCFAIL: Memory allocation of 65536 bytes failed from 0x8011A088, alignment 32 Dec 25 15:44:48 Cisco 69: Pool: Processor Free: 79776 Cause: Memory fragmentation Dec 25 15:44:48 Cisco 70: Alternate Pool: None Free: 0 Cause: No Alternate pool Dec 25 15:44:48 Cisco 71: Dec 25 15:44:48 Cisco 72: -Process= "ARP Input", ipl= 0, pid= 8 Dec 25 15:44:48 Cisco 73: -Traceback= 0x8054EE44 0x800F750C 0x800FD13C

0x8011A08C 0x803A69B4 0x807036A4 0x808F7C30 0x808F8250 0x808F8EDC 0x803B2F54 0x800DDF38 0x800E11F8 Dec 25 15:44:49 Cisco 74: 000065: *Dec 25 09:54:01.476 PCTime: %FIB-3-NOMEM: Malloc Failure, disabling CEF Dec 25 15:44:49 Cisco 75: -Traceback= 0x8054EE44 0x803A69FC 0x807036A4 0x808F7C30 0x808F8250 0x808F8EDC 0x803B2F54 0x800DDF38 0x800E11F8

As you can see, I received a Christmas present yesterday.

I've searched the Cisco support forums and I've googled this error, and I have not found any real solution. When this happens, the router is inac- cessible and has to be reloaded from the serial console or power-cycled.

Can anyone here offer a clue or solution? I am no IOS power user either, so please reply with cogent instructions.

Thanks.

Reply to
VAXman-
Loading thread data ...

AIL: Memory allocation of 65536 bytes failed

fragmentation

0x8011A08C 0x803A69B4 0x807036A4 0x808F7C30 0x808F8250

EM: Malloc Failure, disabling CEF

0x808F7C30 0x808F8250 0x808F8EDC 0x803B2F54 0x800DDF38

Obviously - the router has exhausted its memory.

If you check with sh mem router#sh mem Head Total(b) Used(b) Free(b) Lowest(b) Largest(b) Processor 83099124 70676188 27469044 43207144

40316708 38171256 I/O 7400000 12582912 3925532 8657380 8530336 8537212

and worry about the Processor, free, lowest, largest and also compare how that changes in time against sh proc mem entries per process

you may be able to work out which process is capturing all of the memory.

Since you have already been notified that an ARP process is the one failing the allocation I would check that you are not doing proxy arp on your internet facing interface, or for that matter on th inside interface.

int fa 4 ! If I recall correctly no ip proxy-arp

If none of this makes any sense post

sh proc mem

over several intervals before the router crashes.

Reply to
bod43

I have this on Interface FastEthernet4 as well as Vlan1.

I'll try.

I found one discussion about this and added some arp timeout values that should free up the cached translations, I hope. I set FastEthernet4 to

300 seconds and Vlan1 to 900 seconds. I'll be watching the memory over the next couple of days.

Any other suggestions, I'm all ears at this point having pretty much ex- hausted other paths.

Reply to
VAXman-

I found it. I removed:

# ip route 0.0.0.0 0.0.0.0 FastEthernet4

and replaced it with:

# ip route 0.0.0.0 0.0.0.0 24.xx.yy.zz (IP address up upstream router,)

I removed the timeouts that I had installed too.

# show process memory 8

has shown a rock steady memory use now at 7176 bytes.

Reply to
VAXman-

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.