Re: 20010531: ldm 5.1.3 with RH 7.1 thrashing

Good Morning, Art,

Using a feed from WSI plus a few others along with a 2Gb queue, I tried
again last week to duplicate the zombie rpc.ldmd processes and the
excessive disk I/O on RH7.1, but was unsuccessful.  But, my machine is a
test machine, not a production machine.  I'm guessing that you have a
lot more filing and decoding actions going on then I did.

Here's one more bit of RH7.1 info.  We just got a new machine on which
we installed 7.1.  Being new, it had a SCSI controller that could
transfer at the rate of 160Mbits/sec.  But, we found that the kernal
could not keep up.  This did not appear with "normal" processing, but
appeared when we were doing backups and the disk I/O was intensive. 
With a tape to disk transfer it appeared after 45 minutes to an hour. 
After experimenting, our sys admin solved the problem by changing the
transfer rate of the controller to 80Mbits/sec.

I don't know if this could be a factor for you or not.  I recall that
this was a brand new machine for you, so maybe...

Also, FYI, (maybe I told you this) Gilbert is running 7.1 and also
reports lots of disk I/O with his 600Mb queue.  

I'm sorry I can't help you more with this. We'll keep this info in our
archives and I'll let you know if I discover anything further.  Any info
you can provide us would also be appreciated.

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