Professional Documents
Culture Documents
If the FPC has been restarted to recover.From FPC shell, can you please provide
the following :
Start shell pfe network fpc 2
Show syslog messages
Show nvram
For Error/Interrupt Status (at address 0x10 of i2csc), the customer is using SCBE2:
That I assume shares the same i2c circuitry spec as SCBE, so you can refer to the
following spec :
<-- at Table 3-1 and/or section 3.2.10
http://www-in/eng/cvs_pdf/spec/atlas/fpga/i2cs_scbe/i2cs_design.pdf
So I would think that this is more of a symptom due to the power failure,
represented by bit[4] especially. For now we can stick to "Power Volt Status"
(around 0x2A, 0x24, 0x2E, 0x3A, 0x40) as the significant error symptom.
-----------------------------------------------------------------------------------
-
----------------------
2- Dec. 19 04:57 AM, start to see FRU online messages for RE-0:
> During that time, RE-1 was trying to become master RE, which indicate that there
might be a communication issue between the 2 REs.
3- Later system reported fabric drops to FPC#2 , which triggered FPC to be reset :
http://www.juniper.net/techpubs/en_US/junos12.3/topics/concept/fabric-failures-
corrective-actions-mx-routers.html
{master}
dave@re0.glasbo-rbr1<mailto:dave@re0.glasbo-rbr1>> show chassis fabric reachability
Fabric reachability resolution: Fabric degradation healed after phase Plane and FPC
restart
> Below logs are due to fabric issue between FPC#2 and CB#0
Dec 19 18:35:57 FR: Enqueueing AFPC dest disable event for slot 2
Dec 19 18:35:57 FR: Enqueueing AFPC dest disable event for slot 2
Dec 19 18:35:57 send: yellow alarm set, device CB 0, reason Check CB 0 Fabric Chip
0
Dec 19 18:35:57 CHASSISD_SNMP_TRAP7: SNMP trap generated: fabric plane check
(jnxFruContentsIndex 12, jnxFruL1Index 1, jnxFruL2Index 0, jnxFruL3Index 0,
jnxFruName CB 0, jnxFruType 5, jnxFruSlot 0)
Dec 19 18:35:57 send: yellow alarm set, device CB 0, reason Check CB 0 Fabric Chip
1
Dec 19 18:35:57 CHASSISD_SNMP_TRAP7: SNMP trap generated: fabric plane check
(jnxFruContentsIndex 12, jnxFruL1Index 1, jnxFruL2Index 0, jnxFruL3Index 0,
jnxFruName CB 0, jnxFruType 5, jnxFruSlot 0)
Dec 19 18:35:57 FM: Message rcvd from pb 1 (type 4, subtype 325), pb_up:1
Dec 19 18:35:57 FM: Received plane control ack from PFE board 1, stage:NULL stage,
pb_up:1
Dec 19 18:35:57 FM: plane ctl ack pb 1 toggle_plane_mask:0x00 ...
Dec 19 18:35:57 FM: plane status ...
Dec 19 18:35:57 FM: 2 0 0 0
Based on issue 2 and 3, we might expect an issue with RE0, so we need to perform
the below action plan. Perform RE switchover, after that we need to monitor the
status of RE0 and CB0.Please provide following logs from the FPC : Start shell
pfe network fpc2show nvramshow syslog message During Dec 19, was there any change
in traffic behavior ?After RE switchover, please provide full /var/log from both
REs.
-----------------------------------------------
1)collect the below outputs :
I need following output from both the RE to find out if there is any issue with the
internal RE-to-PFE or RE-to-RE communications
--------------------------------------------
2. I saw the frequent increment in input error counters on em0 interface of RE1
I tried to find out the reason behind the rapid increment in the input error but it
doesn't accounted in the other counters, details can be seen below
3. There are some historic errors that you can see in dmesg output but they are not
increasing. On RE0 these counters all 0
{backup}
imtech@re1.glasbo-rbr1> show tnp connectivity 0x4 count 20
.\x08 \x08.\x08 \x08.\x08 \x08.\x08 \x08.\x08 \x08.\x08 \x08.\x08 \x08.\x08 \x08.\
x08 \x08.\x08 \x08.\x08 \x08.\x08 \x08.\x08 \x08.\x08 \x08.\x08 \x08.\x08 \x08.\
x08 \x08.\x08 \x08.\x08 \x08.\x08 \x0820 of 20 pings
-------------------------------------
Dec 18 13:19:05 send: red alarm set, device FPC 2, reason FPC 2 Hard errors
>> Can we open 2 sessions to the router and try to collect FPC boot
>>messages:
> From the the other session try to restart the FPC.
3- can we have full /var/log from both RE0 and RE1 and upload it to the
case.