Header Only - DO NOT REMOVE - Extreme Networks

Stay away from EXOS 22.2.1.5


Code 22.2.1.5 running on x460G2 has MAJOR stability issues with stacking. I don't suggest running this code in production.

13 replies

Userlevel 2
What are you seeing exactly? I have a few X460-G2's in stacking config running 22.2.1.5.
Haven't had any stability issues.
Log Entries:

03/20/2017 09:35:44.27 Slot-1: Internal error - can't translate ids Slot-6 unit 49

03/20/2017 09:35:44.27 Slot-1: pibConduitMasterRcvOneSlot(): failed, rc=-3, errno=0, slot=6
03/20/2017 09:35:44.02 Slot-2: System Error 0: Conduit receive error encountered
03/20/2017 09:35:44.02 Slot-2: pibConduitMasterRcvOneSlot(): failed, rc=-3, errno=0, slot=6

03/20/2017 09:35:28.11 Slot-6: Process tBcmxSync pid 2087 died with signal 11
>show debug system-dump slot 6
Please wait. Your command is being processed.

===============================================
Slot-6 system dump information
===============================================
core_dump_info storage: 8/5120 used [empty]
failure: process crash
time: Mon Mar 20 09:35:28 2017
process tBcmxSync
pid 2087
signal 11
I think it might be in stacks 6 or bigger.. not 100% sure yet... still working with Extreme on a case.
Userlevel 2
The largest stack i have running is 3 with VIM-2ss.
I would appreciate it if you could post your result with Extreme Networks in here as we also plan to upgrade a stack with 6 switches to 22.2.1.5.

Thanks!
Userlevel 7
Hi All,

I wanted to post on this thread to let the community know that we are currently working through these issues with our development team.

I will be sure to come back to this thread with an update once we have more information on this.
Hello all, I am currently testing various situations with 22.2.1.5 code and also updating Extreme as I discover things. I will post any helpful information when I find it.
Userlevel 4
Hello Guys,

I'm trying the EXOS 22.2.1.5 at my lab (X460-G2 and NetSight 7.0.9.4) and I found an issue (still under investigation, but it vanishes if i downgrade to code 22.1.1.5)...

The issue is related to some alarms arising on NS about "Interface Usage Limit", returning some odd values only in this code version


Best regards,

-Leo
Userlevel 7
Leo wrote:

Hello Guys,

I'm trying the EXOS 22.2.1.5 at my lab (X460-G2 and NetSight 7.0.9.4) and I found an issue (still under investigation, but it vanishes if i downgrade to code 22.1.1.5)...

The issue is related to some alarms arising on NS about "Interface Usage Limit", returning some odd values only in this code version


Best regards,

-Leo

Hi Leonardo, I've forked this into a separate thread so that this doesn't get lost in the other thread, as well as to help keep these two issues better organized. Please reference the new conversation here: Interface Usage Limit returns odd values on EXOS 22.2
Userlevel 4
Leo wrote:

Hello Guys,

I'm trying the EXOS 22.2.1.5 at my lab (X460-G2 and NetSight 7.0.9.4) and I found an issue (still under investigation, but it vanishes if i downgrade to code 22.1.1.5)...

The issue is related to some alarms arising on NS about "Interface Usage Limit", returning some odd values only in this code version


Best regards,

-Leo

Awesome! Thanks!
Userlevel 7
Hi All,

Investigating this with our development team, we've narrowed it down to an issue with diffserv replacement and policy.

The quick explanation is that when both diffserv examination and policy are enabled, some ACLs get corrupted that can lead to a process crash on non-master nodes in a stack. A workaround on 22.2 would be to disable diffserv replacement. A fix is scoped for 22.3.

See the following GTAC Knowledge article for more information:
https://gtacknowledge.extremenetworks.com/articles/Solution/tBcmxSync-process-crash-with-Policy-and-...
Hi all,

I experienced problems with IPv6, local services and 22.2.1.5 firmware. When I upgraded to this version, I could no longer access the switch through ssh or telnet ports using IPv6 address, but only with IPv4 address.

Nmap pointed to IPv6 switch address shows all ports as closed, and to IPv4 address shows ssh and telnet ports open.

Anyone with same issue? I was able to reproduce the same behavior with EXOS-VM_v22.2.1.5-disk1.qcow2 and GNS3.

Best regards,

Ariel
Userlevel 7
alm wrote:

Hi all,

I experienced problems with IPv6, local services and 22.2.1.5 firmware. When I upgraded to this version, I could no longer access the switch through ssh or telnet ports using IPv6 address, but only with IPv4 address.

Nmap pointed to IPv6 switch address shows all ports as closed, and to IPv4 address shows ssh and telnet ports open.

Anyone with same issue? I was able to reproduce the same behavior with EXOS-VM_v22.2.1.5-disk1.qcow2 and GNS3.

Best regards,

Ariel

Hi Ariel, I've split this out to another thread so that it gets better visibility, rather than getting lost under another thread. Please reference the new conversation here: Issues with IPv6 on EXOS 22.2

Reply