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

Re: 20020213: [Fwd: 20020207: missing products]

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: Re: 20020213: [Fwd: 20020207: missing products]
  • Date: Wed, 13 Mar 2002 15:10:19 -0700
Harry Edmon wrote:
> 
> I had to make some changes to /etc/nsswitch.conf - your changes goofed up our
> NFS mounts.
> 

Oops.  Very sorry about that.  We three looked at the fix, and now will
know better.


> I found the DEC Alpha problem.  What is happening is occasionally the write on
> line 439 of ulog/ulog.c is returning a -1 with errno of 55 - No buffer space
> available.  The program then calls closeulog, which closes the log and sets
> logFd to -1.  The following code then does a fork and tries to write to the
> console.  I believe the waitpid at the of this code fails when a signal occurs
> in the server before the child finishes (there seems to be a bunch of SIGCONTs
> going on), and nothing ever goes back and does a wait - thus the zombie
> processes.
> 
> There are probably a number of ways to fix what happens when the write error
> occurs - but I leave that to you.  It also may be possible to increase the
> buffer space on the Alpha for the logging calls.
> 

Thanks very much for tracking this down!  This will be very helpful.


> But, I am now tracking down a different bug in the LDM in Linux - I am trying 
> to
> parallel the setup on air on a Linux box.  When I start up the LDM it tries to
> get data from all the same sources.  Two of the sources (which are schools)
> don't currently have this machine listed as authorized to get data from them.
> But instead of just waiting and trying again the processes asking for the 
> feeds
> are using large amounts of CPU time.  I will let you know if I find this bug.
> 

Again, I really appreciate your effort in uncovering and tracking bugs.

I'm disappointed to report that the nsswitch.conf change does not appear
to have fixed the problem.  At first I was excited because it did not
occur for 8 hours after the change, but after that it occured at least 4
times.  I guess it's time to move to plan C, as soon as I figure out
what that is...


Anne
> --
> Dr. Harry Edmon                 E-MAIL: address@hidden
> 206-543-0547                            address@hidden
> Dept of Atmospheric Sciences    FAX:    206-543-0308
> University of Washington, Box 351640, Seattle, WA 98195-1640

-- 
***************************************************
Anne Wilson                     UCAR Unidata Program            
address@hidden                 P.O. Box 3000
                                  Boulder, CO  80307
----------------------------------------------------
Unidata WWW server       http://www.unidata.ucar.edu/
****************************************************