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

20030321: LDM 6.0.2 problem on Linux 7.3 with large queue

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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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: 20030321: LDM 6.0.2 problem on Linux 7.3 with large queue
  • Date: Fri, 21 Mar 2003 13:11:23 -0700
Arthur,

> To: <address@hidden>
> From: "Arthur A. Person" <address@hidden>
> Subject: LDM 6.0.2 problem
> Organization: Penn State

The above message contained the following:

> Sorry for getting a slow start on installing LDM 6.0...  I was home 2
> weeks while my wife had a baby and this week has been catch-up.

Congratulations!

> I've just tried installing LDM 6.0.2 and couldn't get it to start.  Here's
> the context:
> 
>    LDM 6.0.2
>    RedHat 7.3
>    Built with CFLAGS="-O -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE"
>    New (larger) queue of 6144M (was using 3G queue with LDM 5.2.1)
>    Use "-m 3600" on the rpc.ldmd command line
>    Did ldmadmin start and got one rpc.ldmd process which didn't do much
>    Did ldmadmin watch and after a couple minutes got this:
> 
> Mar 21 14:31:18 pqutil: assertion "pq->rlp->nalloc == pq->nalloc &&
> pq->tqp->nalloc == pq->nalloc && pq->sxp->nalloc == pq->nalloc" failed:
> file "pq.c", line 3905
> Abort (core dumped)
> 
>    In ldmd.log, got:
> 
> Mar 21 14:28:31 ldm rpc.ldmd[28520]: Starting Up (version: 6.0.2; built:
> Mar 21 2003 08:55:52)
> Mar 21 14:30:43 ldm rpc.ldmd[28520]: assertion "pq->rlp->nalloc ==
> pq->nalloc && pq->tqp->nalloc == pq->nalloc && pq->sxp->nalloc ==
> pq->nalloc" failed: file "pq.c", line 3905
> 
> Hopefully, it's not something dumb I'm doing.  Let me know what you think
> is going on and I'll retry when appropriate.

Based on the declaration of the mmap() function on our RedHat 7.3 and
8.0 systems, I don't believe that it's possible to have a product-queue
larger than about 4 gigabytes. The "length" argument of that function is
declared as a "size_t" -- which appears to always be an unsigned, 4 byte
integer.

If you reduce the queue size to around 4 gigabytes, then it should
work.

>                                  Art.
> 
> Arthur A. Person
> Research Assistant, System Administrator
> Penn State Department of Meteorology
> email:  address@hidden, phone:  814-863-1563

Regards,
Steve Emmerson