[evla-sw-discuss] fallout from EVLA power outage

[Kevin Ryan]

We decided that we would purposely not allow the correlator boards to  
power up automatically, after AC mains recovery, for safety's sake.    
A human has to initiate power up.  So it is not critical that cpcc's/ 
mccc recover by themselves.

One advantage of making it so the cpcc's/mccc recover automatically is  
that it would allow the operator to initiate correlator power up (and  
not a team of engineers like this weekend :).  Whether or not we think  
that ability is worth the effort is a decision the policy makers must  
make.  However, given the other problems associated with  power losses  
(signals, timecodes, etc.), we probably aren't ready for a turnkey  
power recovery anyway.

Kevin

On Jul 26, 2010, at 9:24 AM, James Robnett wrote:

>
>   In addition the CPCCs don't power up properly on power
> loss.  They require manual intervention.
>
>   There's been discussion about this in the past,  one school
> of thought is that's the proper behavior, the other is that
> it's not.
>
>   MCCC suffers the same problem.  Both can probably be fixed
> via a BIOS flag.
>
> James
>
>
> Martin Pokorny wrote:
>> Ken Sowinski wrote:
>>> 4.  The CBE had no problems.  I know that James and Martin
>>>     both worked earlier in the day to make it usable.
>>
>> The script /etc/rc3.d/S99widar-cbe didn't run properly according to  
>> the
>> reports. However, that script is not needed by the CBE; it's a java  
>> CBE
>> web server that Kevin created, but it's not being used currently.
>>
>> The CBE software could be started via the "wcbe_services" command.  
>> It's
>> probably worth creating an rc script to run "wcbe_services all  
>> start" as
>> the "wcbe" user on system boot. However, to start writing BDF  
>> files, the
>> Lustre filesystem will have to be mounted by the CBE nodes, and  
>> that's
>> not done by the wcbe_services script.
>>
> _______________________________________________
> evla-sw-discuss mailing list
> evla-sw-discuss at listmgr.cv.nrao.edu
> http://listmgr.cv.nrao.edu/mailman/listinfo/evla-sw-discuss