20050713: McIDAS: Error - resource temporarily unavailable

>From: "Kwan-yin Kong" <address@hidden>
>Organization: CCNY
>Keywords: 200507131709.j6DH91jo022573 McIDAS shared memory pipe read

Hi Kwan,

>      Thanks for your last reply regarding time selection 
>for sub-hour surface data.  I was able to overlaid these 
>data onto radar images.

I'm glad that this worked for you.

>      There is another problem that has come up over the 
>past month or so when I run McIDAS.  The error message I 
>got was "pipe read: resource temporarily unavailable". 
> This error had not occurred before, and it occurs 
>intermitently.  An example of the error message appears 
>below.
>
>IMGDISP RTIMAGES/GE-IR LATLON=32 80 PLACE=C MAG=1 1 
>EU=IR-KONG1.ET SU=X ALL=1 24 SF=YES REFRESH='EG 
>GRA=(GRA);MAP FILE=OUTLUSAM MCOL=1 WID=1 GRA=(GRA);BAR 
>(GRA) SU=IRTEMP X COL=7 ORIENT=VER;TE ?LASTFRM (GRA)'
>Beginning Image Data transfer, bytes= 522896
>IMGDISP: loaded frame  1
>EG GRA=   1 ;MAP FILE=OUTLUSAM MCOL=1 WID=1 GRA=   1 ;BAR 
>   1  SU=IRTEMP X COL=7 ORIENT=VER;TE ?LASTFRM    1
>Erased graphic frame(s) 1-1
>MAP: Completed frame 1
>Beginning Image Data transfer, bytes= 522896
>EU: Restoring IR-KONG1.ET to frame(s)= 1
>EU: Done
>pipe read: Resource temporarily unavailable
>IMGDISP: done
>IMGDISP failed, RC=2
>Done
>?LASTFRM    := 1
>  
>It appears to me that this error occurs when there are 
>more than one McIDAS job running at the same time.  Since 
>I am running cron jobs that execute shell scripts of 
>McIDAS commands hourly, and PC's can be connected to the 
>host machines at any time, there is a good chance that 
>there are more than one McIDAS job running at the same 
>time.  I don't know if that is the source of the error. 
> But again, I have not seen this error prior to this past 
>month or so.

Without having looked at things on your machine, I would
guesss that the problem is that there are a number of shared
memory segments that have not been properly deallocated.
These segments could still exist because of aborted McIDAS
"sessions" (including executions run by script) and by
improperly exited McIDAS XCD processes.  To see if this
is your problem do the following:

<as 'mcidas'>
cd ~mcidas/.mctmp
ls -alt

<as 'ldm'>
cd ~ldm/.mctmp
ls -alt

If you have a lot of directories in each list that are old, you
may have shared memory segments that need to be deallocated.
Deallocation needs to be done by the user that created the
segments (or 'root').  The system commands to use are:

ipcs        <- lists shared memory segments and more
ipcrm       <- removes shared memory segment(s)

The process is to:

** if there are shared memory segments in the 'ldm' account:

- stop the LDM
- remove all subdirectories of ~ldm/.mctmp
- use ipcs to list out all allocated shared memory setments
- use ipcrm to remove those segments

Do the same thing as 'mcidas' or whatever other user is running the
cron-initiated scripts that are now having problems.

By the way, the easiest way to clean up the shared memory segments
is to reboot.  This approach, however, does not take care of the
subdirectories under ~<user>/.mctmp; you will need to remove
those by hand.

Cheers,

Tom
--
NOTE: All email exchanges with Unidata User Support are recorded in the
Unidata inquiry tracking system and then made publicly 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.