Header Only - DO NOT REMOVE - Extreme Networks

SummitStack slot unexpectedly resets


Userlevel 1
Hi all, I got one X460 (slot-1) and one X440 (slot-2) in a stack, connected using standard SummitStack cables in a ring topology. Prior to slot-1 unexpected reboot, I had couple of the following entries in a log: Slot-1: IPv4 multicast entry not added. Hardware Group Table full. Could this slot-1 reset be due to some wrong (default) multicast configuration ? Both switches have the same 16.1.1.4 EXOS image. Regards 🙂 Robert

14 replies

Userlevel 6
Hi Robert,

The unexpected reboot and the log you received are probably unrelated.

As far as the log message you received. The x440 switch was designed as mainly a edge switch so its hardware tables are much smaller than the x460. This limits the stack because they share resources. There are some steps that you can take to help minimize the entries in the tables:

https://gtacknowledge.extremenetworks.com/articles/How_To/Multicast-Entry-not-Added-Hardware-Table-Full

The unexpected reboot we would need to see the output before the unexpected reboot log. If there is nothing this typically means that the switch lost power so I would check the power source and the power cable. If you have a syslog server setup this might have more information than the switch.

Let me know if you have any questions.
Userlevel 1
Hi Patrick,
Userlevel 1
Hi Patrick,
Userlevel 1
Just wanted to add a newline and we have a very short reply 🙂 Delete those two posts above if you can 🙂
Userlevel 6
No problem. It has been done. Did you have any questions about my reply above?
Userlevel 1
Thank you, Patrick Here you have a little longer output: 2015-09-11 05:56:10.22 Slot-2: Slot-2 FAILED (1) Dual Master 2015-09-11 05:56:09.66 Slot-2: Node State[5] = FAIL (Dual Master) 2015-09-11 05:56:07.99 Slot-2: pibConduitMasterRcvOneSlot(): failed, rc=-3, errno=0, slot=1 2015-09-11 05:55:56.51 Slot-2: Slot-1 FAILED (1) 2015-09-11 05:55:56.50 Slot-2: PRIMARY NODE (Slot-1) DOWN 2015-09-10 11:57:03.58 Slot-1: IPv4 multicast entry not added. Hardware Group Table full. 2015-09-10 09:56:16.94 Slot-1: IPv4 multicast entry not added. Hardware Group Table full. 2015-09-10 08:28:08.35 Slot-1: IPv4 multicast entry not added. Hardware Group Table full. 2015-09-10 07:25:33.96 Slot-1: IPv4 multicast entry not added. Hardware Group Table full. 2015-09-10 06:08:31.75 Slot-1: IPv4 multicast entry not added. Hardware Group Table full. 2015-09-09 13:49:40.67 Slot-1: Login failed for user root through ssh (10.1.103.101) 2015-09-09 09:59:57.42 Slot-1: IPv4 multicast entry not added. Hardware Group Table full. 2015-09-09 08:59:45.08 Slot-1: IPv4 multicast entry not added. Hardware Group Table full. 2015-09-09 07:59:03.23 Slot-1: IPv4 multicast entry not added. Hardware Group Table full. 2015-09-09 06:51:18.25 Slot-1: Login failed for user root through xml (10.1.103.101) I'll check the article you mentioned and the power conditions.
Userlevel 6
I am not sure which slot this output was taken from but can you gather the same output from the opposite slot. You can do this through the master and type in "telnet slot , login, then run the command accordingly.
Userlevel 3
look at the stacking ports for errors - ( show port stack-port rxerrors)

I had a switch rebooting and it seemed to be related to a bad stacking cable. I could see the rx crc error count climbing every second and getting extremely high then the slot would reboot...

I replaced the stacking cable and the errors stopped and the stack stayed up. - just a thought

Dave
Userlevel 1
So far I can say that it is not the cable issue: there are no rxerrors. GTAC case is open.
Userlevel 1
Your x460s are linked to other equipements via XGM3-2sf module ?
Userlevel 1
No, it is linked through SummitStack module, although XGM3S-2sf module is inserted in X460 (but not utilized).
Userlevel 1
Update: following GTAC recomendation, we updated stack to EXOS version 16.1.1.4-patch1-1; waiting for any feedback from customer.
Userlevel 1
No info about stack resets - case closed.
Userlevel 7
Robert Zdziebło wrote:

No info about stack resets - case closed.

Thanks for coming back to update the thread, Robert!
I'll mark this as "answered."

Reply