[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20040615: Weather3.admin.niu.edu update

>From: Gilbert Sebenste <address@hidden>
>Organization:  NIU
>Keywords:  200406150354.i5F3sOtK011267 LDM Fedora Core Linux

Gilbert,

Kevin said:
>> I'd say that there is a reasonable chance that is what was causing your
>> problems.  It is a good explanation for the SIGBUS crashes.  It *could* be
>> an explanation for your corrupt filesystem problems, though you said your
>> system logged I/O errors.  If you get any more I/O errors from you disk
>> drive, I'd replace it.
>> 
>>      Kevin W. Thomas
>>      Center for Analysis and Prediction of Storms
>>      University of Oklahoma
>>      Norman, Oklahoma
>>      Email:  address@hidden

>I am cc'ing this to support. Since Tom Yoksas mentioned that the LDM does 
>a memory map,

Right.  The LDM queue is a memory mapped file.

>I would suppose that this would cause a corrupted disk as 
>well. Tom, I didn't explain clearly, but before this happened, I was also 
>getting corrupt disk errors that required a e2fsck.

I think I remember seeing this.  At the time, it made sense to think
that this was being caused by a bad disk drive.

>There were blocks that 
>were duplicated, but no apparent failure of the drive.

Duplicated...  Interesting.

>This problem has 
>not worsened, so I suspected something else. What say you?

I can't constructe a mental model of how your improperly seated DIMM
could cause file system corruption, but this is not unimaginable
especially given your comments about having to keep your LDM queue
small.  I say that time will tell :-)

>******************************************************************************
> *
>Gilbert Sebenste                                                     ********
>(My opinions only!)                                                  ******
>Staff Meteorologist, Northern Illinois University                      ****
>E-mail: address@hidden                                               ***
>web: http://weather.admin.niu.edu                                      **
>Work phone: 815-753-5492                                                *
>******************************************************************************
> *
>
--
NOTE: All email exchanges with Unidata User Support are recorded in the
Unidata inquiry tracking system and then made publically available
through the web.  If you do not want to have your interactions made
available in this way, you must let us know in each email you send to us.

>From address@hidden  Tue Jun 15 16:09:32 2004

Gilbert...

Memory mapping *is* the way that a lot of things work behind your back though I
don't see how this could cause a filesystem corruption problem.  If a memory
problem caused a driver to misbehave, I'd also expect frequent kernel panics.

Duplicate blocks are a symptom of a malfunctioning disk drive.

I'd consider replacing the drive if you see the hint of another problem with
it.

        Kevin W. Thomas
        Center for Analysis and Prediction of Storms
        University of Oklahoma
        Norman, Oklahoma
        Email:  address@hidden

>From address@hidden  Tue Jun 15 23:56:50 2004

> I'd consider replacing the drive if you see the hint of another problem with
> it.

OK, thanks. It's still under warranty. I'm watching it close!

*******************************************************************************
Gilbert Sebenste                                                     ********
(My opinions only!)                                                  ******
Staff Meteorologist, Northern Illinois University                      ****
E-mail: address@hidden                                               ***
web: http://weather.admin.niu.edu                                      **
Work phone: 815-753-5492                                                *
*******************************************************************************