8600 boots to backup cpu ?

[gillymonst]

Hi,
What would cause a new 8600 to roll over to the backup cpu in the middle of a sunday night w/ no traffic load. Here's the log file a few lines before cpu #5 drops and boots to redundant cpu #6. Not sure what "Killing task 17de258 after timeout" means but the rest of the log file shows a healthy boot to cpu #6. Oh yea, this is 4.0.3.2 code.

Thanks for any insight.


<NP> 64:e00b73262be8dda95341096a9dd5f73cf9f6b8b322bb73840a2bd861cf5e042beca5f31942e920ecab88809603e056e1d1ffd667578494ed59785c295a49ecdf</NP> CPU6 [06/22/06 16:18:58] HW INFO System activity performed
<NP>000:</NP> [08/20/06 16:13:40] The previous message repeated 59 time(s).
<NP> 24:e00b73262be8dda921c7998c0da8c509e1678b41471f0cd5</NP> CPU6 [08/21/06 00:26:31] HW INFO Stand-by CPU in slot # 6 becoming master...
<NP> 24:261340581186ac6bca964417d3fdbea827e5446d5f33fd56</NP> CPU6 [08/21/06 00:26:34] SW INFO Killing task 17de258 after timeout
<NP> 24:e00b73262be8dda921c7998c0da8c509e1678b41471f0cd5</NP> CPU6 [08/21/06 00:26:41] HW INFO Card inserted: Slot=6 Type=8691SF
<NP> 24:e00b73262be8dda921c7998c0da8c509e1678b41471f0cd5</NP> CPU6 [08/21/06 00:26:41] HW INFO Card inserted: Slot=1 Type=8608GBM
<NP> 24:e00b73262be8dda921c7998c0da8c509e1678b41471f0cd5</NP> CPU6 [08/21/06 00:26:41] HW INFO Card inserted: Slot=2 Type=8608GBM
<NP> 24:e00b73262be8dda921c7998c0da8c509e1678b41471f0cd5</NP> CPU6 [08/21/06 00:26:41] HW INFO Card inserted: Slot=5 Type=8691SF
<NP> 24:e00b73262be8dda921c7998c0da8c509e1678b41471f0cd5</NP> CPU6 [08/21/06 00:26:41] HW INFO Card inserted: Slot=8 Type=8648TXM
<NP> 24:e00b73262be8dda921c7998c0da8c509e1678b41471f0cd5</NP> CPU6 [08/21/06 00:26:41] HW INFO Card inserted: Slot=9 Type=8648TXM

 

[anthonyanderberg]

I'd assume 17de258 is the process ID number of something important, and that that process stopped responding to the hardware watchdog and so a failover was triggered.

It could be a fluke, if it happens again I'd call Nortel and send them the log files - and I'd search your log files to make sure it hasn't happened in the past without your noticing. The Nortel folks have got commands that show more detail than our normal CLI. Having said that, they'll probably start by suggesting you upgrade microcode - I had several problems with code in the 4.0.x branch prior to release 4.0.6