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

20010918: reboot of motherlode.ucar.edu scheduled for 18Z

>From: "Robert Mullenax" <address@hidden>
>Organization: NMSU/NSBF, Universal Weather
>Keywords: 200109181909.f8IJ9T113502 McIDAS Solaris shared memory

Robert,

>Just curious..Why are you having to increase the amount of shared
>memory?

We had a situation that makes absolutely no sense.

motherlode had been up for 134 days without a reboot.  Match that,
Linux :-).  Actually, while on the subject of Linux vs Solaris, I would
like to put my pitch in for FreeBSD.  It is a LOT faster than either
Linux or Solaris x86, and it is incredibly stable.  I _know_ that you
would get agreement on this point from Jim Koermer of Plymouth State
College!!

During the 134 days of operation, McIDAS-XCD was merrily decoding data,
and McIDAS ADDE was serving that data to a variety of sites across the
country.  Yesterday, at around 22Z all McIDAS-XCD decoding except GRIB
(DMGRID) failed.

All McIDAS decoders except DMGRID periodically exit and are restarted
by the XCD supervisory routine, startxcd.k.  For some _unknown_ reason,
those decoders (data monitors actually) could no longer allocate shared
memory segments which comprise McIDAS User Common, so they could not be
restarted.  A quick check of /etc/system shows that the entries that we
HAD put in there to increase shared memory to 512 MB had vanished.  The
mystery is that the timestamp on /etc/system predated the last reboot
of motherlode!  This means that McIDAS-XCD and ADDE routines should
have _not_ worked since the amount of shared memory on the system was
only 1 MB!

>What are you increasing it to?

I increased shared memory from the default 1 MB to the 512 MB recommended
for Sun Solaris systems in:

http://www.unidata.ucar.edu/packages/mcidas/780/mcx/workstation.html

The only reason I included a short comment in my announcement was to
clarify the comment that Anne had made in her earlier announcement.

After modifying /etc/system and rebooting, motherlode is once again
happily running McIDAS-XCD decoders and serving ADDE data.

>Thanks,

Talk to you later...

>Robert Mullenax

 >>From: Unidata Support <address@hidden>
 >>Reply-To: address@hidden
 >>To: address@hidden, address@hidden
 >>CC: address@hidden
 >>Subject: 20010918: reboot of motherlode.ucar.edu scheduled for 18Z
 >>Date: Tue, 18 Sep 2001 12:12:33 -0600
 >>
 >> >From: Unidata User Support <address@hidden>
 >> >Organization: Unidata Program Center/UCAR
 >> >Keywords: 200109181727.f8IHRW108738 IDD motherlode reboot
 >>
 >>LDM Users:
 >>
 >>We have scheduled a reboot of motherlode.ucar.edu, our main IDD
 >>injection node, for 19:00Z today (13:00 MDT time).
 >>
 >>We not sure how long the reboot will take, but we do not anticipate any
 >>problems.  Downstream sites who are concerned about receiving data
 >>during the reboot may want to fail over.
 >>
 >>'motherlode' is being rebooted to reconfigure (increase) the amount of
 >>shared memory that is available on the system.  After the reboot,
 >>McIDAS-XCD decoding and ADDE serving of data should be restored.
 >>
 >>Again, we are sorry for any inconvenience that the reboot may cause.
 >>
 >>Tom Yoksas

Tom