Switching

Expand all | Collapse all

QFX10k Major alarm FPC 0 Errors

  • 1.  QFX10k Major alarm FPC 0 Errors

    Posted 10-22-2018 01:19

    Hi all;

    We have QFX10002-36q running 15.1X53-D62.5 version, we some strange logs and major alarm on FPC 0 as you can see below, have you already seen this issue, how to solve it ? is it a juniper bug ?

     

    thanks for your replies

     

    ---------------------------------------------------------------------------------------------------------------

    qfx10k-1> show chassis alarms

    1 alarms currently active
    Alarm time Class Description
    2018-10-21 06:12:31 CEST Major FPC 0 Major Errors

    ----------------------------------------------------------------------------------------------------------------

    qfx10k-1> show log messages.5.gz | except bgp_listen_accept

    Oct 21 06:12:31 qfx10k-1 fpc0 PE Chip:PE-1[1]: HMCIF: Link4: HMC Fatal Error cmd:62 lng:1 ltag:2 dinv:0 errstat:127 err_cnt:0x40000000
    Oct 21 06:12:31 qfx10k-1 fpc0 PE Chip:PE-1[1]: HMCIF: Link5: HMC Fatal Error cmd:62 lng:1 ltag:2 dinv:0 errstat:127 err_cnt:0x40000000
    Oct 21 06:12:31 qfx10k-1 alarmd[3599]: Alarm set: FPC color=RED, class=CHASSIS, reason=FPC 0 Major Errors
    Oct 21 06:12:31 qfx10k-1 craftd[3600]: Receive FX craftd set alarm message: color: 1 class: 100 object: 104 slot: 0 silent: 0 short_reason=FPC 0 Major Errors long_reason=FPC 0 Major Errors id=150995048 reason=150994944
    Oct 21 06:12:31 tls00-1-q10k craftd[3600]: Major alarm set, FPC 0 Major Errors

    ---------------------------------------------------------------------------------------------------------------



  • 2.  RE: QFX10k Major alarm FPC 0 Errors

    Posted 10-22-2018 01:35

    If you cannot find a PR matching these errors I would suspect HW-failure/error. My memory tells me that HMC refers to the "hyper/hybrid memory cubes" used on this platform.

     

    Please raise a ticket with JTAC to validate this.



  • 3.  RE: QFX10k Major alarm FPC 0 Errors

    Posted 10-22-2018 04:46

    Hi Jonashauge

     

    Thanks for your reply, exactly, this alarm is related to Hybrid Memory Cubes used on QFX platform

    I think this problem is similar to this https://prsearch.juniper.net/InfoCenter/index?page=prcontent&id=PR1300180

    What do you think ?

     

    Best regards;


    #QFX10002
    #HMC


  • 4.  RE: QFX10k Major alarm FPC 0 Errors

     
    Posted 10-22-2018 16:49

    JTAC could verify it conclusively but it does seem similar.  If you don't want to go through the verification process with JTAC you could upgrade to the fixed version listed in the PR and see it clears.

     



  • 5.  RE: QFX10k Major alarm FPC 0 Errors

    Posted 10-29-2018 09:52

    Hello,

    I've tried to upgrade our QFX from 15.1X53-D62.5 to 17.3R3-S1.5, but after  upgrading it was unable to ssh the box, by console it was OK but I saw some errors on log message like:

    root> show system license 
    error: the license-service subsystem is not running

    Despite of this error, I observed all IP traffic was handled by the box, bgp session were UP.

    When I tried to restart sshd I got this:

    root> start shell 
    root@:RE:0% /usr/sbin/sshd 
    /var/etc//sshd_conf: No such file or directory
    root@:RE:0% 

    Then I have some questions may be you know the answers:

     

    1)- What is the path upgrade if you want to go from 15.1X53-D62.5 to 17.3R3-S1.5 ?

    Note that 17.3R3-S1.5 is the recommanded junos version for QFX10002

     

    2)Junos name has changed before the file name was named as: jinstall-host-qfx-10-f-15.1X53-D62.5-domestic-signed.tgz, for 17.3R3 from juniper web site files are named:

    - QFX 10K Fixed Series Switch with Enhanced Automation

    jinstall-host-qfx-10-f-flex-x86-64-17.3R3-S1.5-secure-signed.tgz

     

    - Limited - QFX 10K Fixed Series Switch

    jinstall-host-qfx-10-f-x86-64-17.3R3-S1.5-secure-limited-signed.tgz

     

    - QFX 10K Fixed Series Switch

    jinstall-host-qfx-10-f-x86-64-17.3R3-S1.5-secure-signed.tgz

     

    Which file not contain ssh deamon ?

     

    Based on the error message related to license et not access to the box by ssh, I have rollback to the previous version 15.1X53, after rollback no error on license, ssh to the box OK and the error that push me to upgrade disappear

    root> show chassis alarms
    No alarms currently active

     

    before I have this when I do show chassis alarms

    QFX10k Major alarm  FPC 0 Errors

     

    Thanks in advance for your replies.

    Best regards,



  • 6.  RE: QFX10k Major alarm FPC 0 Errors

     
    Posted 10-29-2018 13:59

    The HMC errors will very likely go away with a reboot, which is what occurred when you changed the code.  As for SSH/non-SSH code, the image with wording "Limited" should be the only one without encryption included.

     

    Unless you plan to use EVPN/VXLAN I would stay on 15.1X53, vs 17.3, at this time.

     

    Hope this helps



  • 7.  RE: QFX10k Major alarm FPC 0 Errors

    Posted 10-29-2018 17:06
    hello. thanks for your reply
    is there any path upgrade to respect in order to upgrade from 15.1X53, vs 17.3?

    Best regards,