A crashing/freezing MTP TIM/TIM displays glibc messages in MTP/TIM and TIM stdout logs.

Document ID : KB000012301
Last Modified Date : 14/02/2018
Show Technical Document Details
Introduction:

  This problem has occurred several times and needed to be documented with next steps.

Question:

   My TIM/MTP TIM is freezing or having restart issues. Looking at the stdout logs, I see messages like:

  *** glibc detected *** timworker: double free or corruption (!prev): 0x00007febc8509d60 ***

 

   OR  

 

   *** glibc detected *** /opt/tim/CA/APM/tim/bin/tim: corrupted double-linked list: 0x00007f15044ab2f0 *** 

   I see nothing in the documentation on this. What should I do on this?

Environment:
All APM 10.x TIMs
Answer:

  These memory-oriented messages are related to the TIM being overloaded. Do the following if these conditions occur:

  1. Reduce traffic to TIM

     a)     First Filter on switch to minimize what it sends to TIM

     b)  Filter as much as possible on MTP HW Filters. There are limitations on the number of arguments

     c)  And always last, filter on Web Server filters

 

  2. Increase workers to "match number of CPUs - 1" and review balancer.cnf settings.

      Example: You have 8 CPU's so 8 - 1 would be 7.  Therefore you would increase the number of workers by 7.

  3. Check protocolstats logs if packets are being dropped/aged out.

  4. Increase TIM watchdog memory