ASR920_SYS-2-MALLOCFAIL_and_SYS-3-CPUHOG

[faruk.zaimovic]

Does anyone know the explanation for the logs that appear on the ASR920 and ME3600? Logs below.

Jun 10 18:27:08.771: %SYS-2-MALLOCFAIL: Memory allocation of 24835080 bytes failed from 0x48FF85C, alignment 0
Pool: Processor Free: 352781344 Cause: Memory fragmentation
Alternate Pool: None Free: 0 Cause: No Alternate pool
-Process= "OSPF-8000 Router", ipl= 0, pid= 389
-Traceback= 7A6984z 1E0E64Cz 1E18D08z 4BCA3A0z 48FF860z 48D9E98z 48FF0A8z 4C72910z 15EB2E8z 15915BCz 1595F50z 1621B68z 1599F38z 1586FACz 158C47Cz 3F4CB28z
Jun 11 01:33:18.236: %SYS-2-MALLOCFAIL: Memory allocation of 24904072 bytes failed from 0x48FF85C, alignment 0
Pool: Processor Free: 352029784 Cause: Memory fragmentation
Alternate Pool: None Free: 0 Cause: No Alternate pool
-Process= "OSPF-8000 Router", ipl= 0, pid= 389
-Traceback= 7A6984z 1E0E64Cz 1E18D08z 4BCA3A0z 48FF860z 48D9E98z 48FF0A8z 4C72910z 15EB2E8z 15915BCz 1595F50z 1621B68z 1599F38z 1586FACz 158C47Cz 3F4CB28z
Jun 11 08:13:04.136: %SYS-3-CPUHOG: Task is running for (2000)msecs, more than (2000)msecs (0/0),process = OSPF-8000 Router.
-Traceback= 0x1E0D470z 0x1E11BC8z 0x4BCA350z 0x48FF7ECz 0x48D9E70z 0x48FF0A8z 0x4C72910z 0x15EB2E8z 0x15915BCz 0x1595F50z 0x1621B68z 0x1599F38z 0x1586FACz 0x158C47Cz 0x3F4CB28z 0x3F47D44z

Jun 9 07:01:06.728: %SYS-3-CPUHOG: Task is running for (2000)msecs, more than (2000)msecs (0/0),process = OSPF-8000 Router.

[marce1000]

 - Check software version being used on the involved platforms, use and or upgrade to an advisory release, if applicable, and or if current versions are getting dated , check if the problems remain afterwards.

 M.

[faruk.zaimovic]

Thank you. 

We have that logs on different versions, and on more devices. 

[Giuseppe Larosa]

Hello @faruk.zaimovic ,

how big is your OSPF 8000 database ?

post a

show ip ospf 8000 database summary.

The first error means a failure in allocating memory to be used by process OSPF 8000.

The second error CPUHOG means the process OSPF 8000 used the main CPU for a too long time more then the threshold of 2000 msecs.

Edit:

in case of a big OSPF database you should consider moving the affected devices in NSSA areas so that you reduce the size of DB on them stll allowing them to act as ASBR if needed now or in the future.

Hope to help

Giuseppe

[faruk.zaimovic]

Hello Giuseppe,

Thanks for your help.

I can not send u show ip ospf database summary, but can show ip route summary. if it can help.

 

[Giuseppe Larosa]

Hello @faruk.zaimovic ,

the number of OSPF routes is not very high a little less then 2,000.

At this point I agree that you should open a Cisco SR with TAC.

Hope to help

Giuseppe

[Richard Burts]

The first message "%SYS-2-MALLOCFAIL: Memory allocation of 24835080 bytes failed" indicates a problem with memory and is frequently the result of a memory leak (bug in the code).

The suggestion of code issues is also supported by the traceback messages in the log. The traceback is always the result of a software issue (problem is not hardware).

If you have a support contract I agree that opening a case with TAC is the thing to do. And I certainly expect them to suggest different version of code.

[johnlloyd_13]

hi,

you'll need to raise a ticket with TAC for cases like this.

they'll usually recommend an IOS upgrade to resolve this kind of issue.