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

Re: 20001107: LDM problem at University of Puerto Rico

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: 20001107: LDM problem at University of Puerto Rico
  • Date: Tue, 07 Nov 2000 11:01:10 -0700
>From: McIDAS <address@hidden>
>Organization: McIDAS proyect
>Keywords: 200011070833.eA78XG409393 LDM mmap IRIX64

Hi Karli,

> Nov 07 00:00:24 3Q:breeze sysu1[16296]: mmap: 64000000 0 1069875200: Not 
> enough space
> Nov 07 00:00:24 3Q:breeze sysu1[16296]: Remap failed. Abandon all hope.
> Nov 07 00:00:24 3Q:breeze sysu1[16296]: pq_last: seq:Not enough space (errno 
> = 12)
> Nov 07 00:00:24 5Q:breeze sysu1[16296]: hiya: 20001106230024.053 TS_ENDT 
> {{WSI,  ".*"}}
> Nov 07 00:00:34 5Q:breeze sysu1[16296]: Growing data by 33980416
> Nov 07 00:00:34 3Q:breeze sysu1[16296]: mmap: 40000000 0 1103855616: Not 
> enough space
> Nov 07 00:00:34 3Q:breeze sysu1[16296]: comings: pqe_new: Not enough space
> Nov 07 00:00:34 3Q:breeze sysu1[16296]:        : 
> 7d2c6ca2f497a3b4aa283edaa7c69a45    16086 20001107000034.115     WSI 001  
> NEX/JUA/SRMV1/200011062355
> Nov 07 00:00:44 5Q:breeze sysu1[16296]: Growing data by 33980416

I like the "Abandon all hope." message :-).

But what this means is that the "pqexpire" program isn't doing its job
of keeping the queue small by deleting old products.  This could be
either because

 - pqexpire is no longer running.  In this case on IRIX64 platforms
   the queue will just keep growing until it uses up all available
   disk space.

 - pqexpire is too slow to keep up with deleting products from the
   queue, in which case the queue will also keep growing until it uses
   up all disk space.

 - pqexpire isn't being invoked often enough to keep the queue from
   growing too large.

Unfortunately the pqexpire program gets slower and slower as the queue
gets bigger and bigger, so if the queue grows beyond a certain size,
pqexpire won't be able to keep up and the downward death spiral you
see in the log files could be the result.

There are a couple of workarounds for this problem:

 - Rebuild the LDM from source, first setting the CPPFLAGS environment
   variable to contain the string "-D_NOGROW" to keep the queue from
   ever growing larger than the size it is created with.

 - Get the current 5.1.2 version of the LDM, which eliminates growable
   queues and also eliminates the need to run the "pqexpire" program.
   We have a binary installation of this available from 

    ftp://ftp.unidata.ucar.edu/pub/binary/irix64_6.5-mips/ldm-5.1.2.tar.Z

   and the installation instructions are available from

    http://www.unidata.ucar.edu/packages/ldm/ldmBinaryInstallList.html

--Russ

_____________________________________________________________________

Russ Rew                                         UCAR Unidata Program
address@hidden                     http://www.unidata.ucar.edu