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

20040609: moving feeds off of (new)thelma



>From: Unidata User Support <address@hidden>
>Organization: Unidata Program Center/UCAR
>Keywords:

Mike,

After we talked, I logged onto (new)thelma and took a look at who
was feeding from it.  I found that there were three sites requesting
multiple feeds from thelma that I had 'ldm' logins to:

rossby.met.sjsu.edu
rainbow.al.noaa.gov
papagayo.unl.edu

I switched all feed from thelma to emo on these machines.  When
I started the reassignment, the load average was 161+.  After all
three had moved off, it dropped to 139 quickly and then to ~105.
It may well be the case that enough load has been moved off of
(new)thelma that I do not have to resort to excluding the CRAFT
and CONDUIT feeds.

I will continue to monitor the load on (new)thelma to see if I
need to drop the CRAFT or CONDUIT feed requests.  If the load
doesn't drop to more reasonable levels, I will probably dump
the CRAFT feed first since it has 8 requests to OU machines.
The next thing I would try is dumping the 5 ALTERNATE CONDUIT
feeds from tgsv32 and leave the single request to atm.

Maybe more to come...

Tom
**************************************************************************** <
Unidata User Support                                    UCAR Unidata Program <
(303)497-8643                                                  P.O. Box 3000 <
address@hidden                                   Boulder, CO 80307 <
---------------------------------------------------------------------------- <
Unidata WWW Service              http://my.unidata.ucar.edu/content/support  <
**************************************************************************** <