Routing
Routing

PFE kernel core dump on EX4200 after enabling IPv6

05.05.17   |  
‎05-05-2017 11:07 AM

Hi

 

We have two EX4200 VC configured as L3 Aggregation and it's working properly with IPv4.

 

We are using the following version:

 

Model: ex4200-24t
Junos: 15.1R5.5

 

Last night we implemented IPv6 on these box and this morning the traffic was stopped and going back (intermittent) for some minutes.

 

So, looking in the log messages, it got the following error:

 

May  5 12:26:01  sw20 init: packet-forwarding-engine (PID 18105) terminated by signal number 11. Core dumped!
May  5 12:26:01  sw20 /kernel: peer_input_pending_internal:[4709] VKS0 for peer type 25 indx 1 reported a sb_state 32 = SBS_CANTRCVMORE
May  5 12:26:01  sw20 /kernel: peer_inputs:4959 VKS0 closing connection peer type 25 indx 1 err 5
May  5 12:26:01  sw20 /kernel: peer_input_pending_internal:[4709] VKS0 for peer type 25 indx 0 reported a sb_state 32 = SBS_CANTRCVMORE
May  5 12:26:01  sw20 /kernel: peer_inputs:4959 VKS0 closing connection peer type 25 indx 0 err 5
May  5 12:26:01  sw20 /kernel: mpfe_mmap pfem mpfe mmap
May  5 12:26:01  sw20 init: Dump Command: /bin/sh (PID 18502) started
May  5 12:26:01  sw20 init: packet-forwarding-engine (PID 18503) started

 

And when this happens, the CPU got almost 100%

 

I didn't any special configuration for IPv6, I configured just like:

 

[edit interfaces ae0 unit 0]
-      family inet6 {
-          address fe80::2/64;
-      }
[edit interfaces vlan unit 10]
-      family inet6 {
-          address 2001:DB8:1::1/48; <-- Changed instead our prefixes
-      }
[edit routing-options]
-   rib inet6.0 {
-       static {
-           route ::/0 {
-               qualified-next-hop fe80::1 {
-                   interface ae0.0;
-               }
-           }
-       }
-   }

 

Have anyone experienced anything similar it ?

 

Thank you

 

 

 

5 REPLIES
Routing

Re: PFE kernel core dump on EX4200 after enabling IPv6

05.05.17   |  
‎05-05-2017 11:12 AM

I forgot to mention that after rollback the environment was normalized.

Highlighted
Routing

Re: PFE kernel core dump on EX4200 after enabling IPv6

05.05.17   |  
‎05-05-2017 11:48 AM
init: packet-forwarding-engine (PID 18105) terminated by signal number 11. Core dumped!

Do the show system core-dump , and if there is file there then I think you need to report this issue to JTAC and let them investigate the coredump file .

regarding this line , signal number 11 means segmentation violation:
https://www.freebsd.org/cgi/man.cgi?query=signal&sektion=3

Regards,
A'bed AL-R.
[JNCSP-SEC JNCDA JNCIS-ENT Ingenious Champion|Sec]
https://srxtech.wordpress.com
Routing

Re: PFE kernel core dump on EX4200 after enabling IPv6

05.08.17   |  
‎05-08-2017 07:33 AM

Hi Abed

 

Thank you for your response.

 

Yes. There is core dump files. Do you know any program to read these core dump ? I tried with lldb but I can't see any special information unless sigsegv signal, like:

 

(lldb) bt all  
* thread #1: tid = 100125, 0x01f54534, stop reason = signal SIGSEGV
  * frame #0: 0x01f54534
    frame #1: 0x01f544e4

 

What I didn't understanding so far is why a simple config got a core dump.

 

Do you have any idea how I could figure out this issue ?

 

Thank you

 

 

Routing

Re: PFE kernel core dump on EX4200 after enabling IPv6

05.08.17   |  
‎05-08-2017 07:43 AM

At this point, you need to engage JTAC to find the root cause of the core-dump.

 

No external tool will give you this information.


=====

If this worked for you please flag my post as an "Accepted Solution" so others can benefit. A kudo would be cool if you think I earned it.

Routing

Re: PFE kernel core dump on EX4200 after enabling IPv6

05.08.17   |  
‎05-08-2017 09:21 AM

Hi

 

Regarding the core dump,  I think that it is relationed by the high cpu usage and it may be relationed with NDP or anything similar on IPv6 that may be flooding the network.

 

Thank you