The following message intended for the ldm-users email list bounced
(non-member address submission). We are forwarding it along...
------- Forwarded Message
Content-Type: text/plain; format=flowed
X-OriginalArrivalTime: 18 Sep 2001 22:40:39.0601 (UTC)
It was your mention of kernel faults and crashes which prompted
my e-mail. I have found that so few people know about Solaris
Intel, that I have taken it upon myself to at least get the word
out. I certainly agree an archive is needed, but I am just
not sure how practical a feed archive is. I would think some sort
of large community machine that archived radar, satellite, model
data, obs, u/a would be pretty handy. We(still work there part time) have
an upper air/ surface archive at NMSU/NSBF and I have responded to data
requests before. Anyway, NMSU/NSBF will help if possible
as a surface/ua data archive. We are too low on the IDD chain
to ensure that we get all of the data though.
>From: Daryl Herzmann <akrherz@xxxxxxxxxxx>
>Subject: RE: LDM Archive Machine
>Date: Tue, 18 Sep 2001 15:26:40 -0500 (CDT)
> My original post was not intended to start a holy war of Linux vs
>Solaris. We have been down that road and know which side IBM is puttting
>1 Billion Dollars into. Hehe :)
> Anyway, I realize that if LDM servers would stay on-line and
>internet connections would stay on-line and people would not run IIS which
>take down OC3 lines, we would not need an archive. There are reasons
>other than Linux stability that make an archive needed in my opinion.
>Your milege may vary...
>On Tue, 18 Sep 2001, Arthur A. Person wrote:
> >On Tue, 18 Sep 2001, Robert Mullenax wrote:
> >> Daryl,
> >> I wish you luck in your endeavor, but IMO, and based on my tests and
> >> observing the e-mails about LDM through the years, the problems with
> >> LDM crashing/kernel problems..etc is due to Linux itself. It certainly
> >> has the well known problem of crashing under heavy load, whereas
> >> Solaris Intel is virtually bulletproof. Not to pick on Gilbert,
> >> he's been a great help to me over the years, but his LDM just crashed
> >> again..he hates Solaris but I am 99.9% sure these bouts he has would
> >> go away if he'd switch.
> >Linux may be a little green yet, but, in my opinion, it's poised soon to
> >be the OS of choice for server systems. Our ldm relay system (no
> >decoding) is RedHat 7.1 with LDM 5.1.3 and it NEVER quits. Runs
> >flawlessly (ignoring configuration issues). I also have a second system
> >running RedHat 7.0 that does decoding and file serving. It too runs
> >nearly flawlessly (I can think of one case where it had a problem a
> >months ago that required a reboot, but it's also 7.0... I need to
> >upgrade). I'm sure Gilbert's problems will be found to be some subtle
> >configuration issue or perhaps a hardware issue, but I don't believe
> >is fundamentally the problem. Regarding Solaris, at this point I'd never
> >go back. Solaris is solid but, in my opinion, it's falling behind in
> >development and overall utility.
> >My 2 cents.
> > Art.
> >> At my previous job at NMSU/NSBF we have been
> >> running
> >> the LDM plus all of the GEMPAK AND McIDAS decoders plus a web server
> >> GARP/McIDAS sessions on a single CPU Gateway PC running Solaris for
> >> years now. The
> >> only problems that have ever occurred were related to hung gplt
> >> caused by my CGI scripts. These occasional to frequent crashes that
> >> LDM-Linux
> >> boxes all seem to have simply don't occur or rarely occur with Solaris
> >> Intel.
> >> Performance is fine as well..and as CPU usage approaches 100% you'll
> >> Solaris
> >> surpass Linux.
> >> Anne can correct me on this but Unidata is running an incredible amount
> >> stuff on their dual PIII Solaris machines and have never had a problem.
> >> Linux has it's place, but with this particular application (the
> >> LDM/decoders)
> >> it simply fails more often than it should. If you continue to have
> >> I would suggest that you at least look into Solaris x86. Solaris will
> >> on the Athlon, depending on the particular motherboard. We had a
> >> problem at NMSU/NSBF which caused us always to lose a bunch of data. I
> >> eventually
> >> arranged for our feed to set his ldm to 5400 seconds as I did mine to
> >> down on that a bit..then when I needed data I put out a call on this
> >> list or the needdata list. There are many people archiving out there.
> >> In any event, good luck, to you.
> >> /**
> >> * Daryl Herzmann (akrherz@xxxxxxxxxxx)
> >> * Program Assistant -- Iowa Environmental Mesonet
> >> * http://mesonet.agron.iastate.edu
> >> */
> >Arthur A. Person
> >Research Assistant, System Administrator
> >Penn State Department of Meteorology
> >email: person@xxxxxxxxxxxxxxxxxx, phone: 814-863-1563
> * Daryl Herzmann (akrherz@xxxxxxxxxxx)
> * Program Assistant -- Iowa Environmental Mesonet
> * http://mesonet.agron.iastate.edu
Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp
------- End of Forwarded Message