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

19990930: McIDAS: Navigation Problems

>From: Michael Keables <address@hidden>
>Organization: DU
>Keywords: 199909301852.MAA26441 McIDAS

Mike,

>When running McIDAS under a user account on cyclone, I can't seem to be
>able to get the maps to overlay on the images. When I run mcidas as a
>user, the first message I get 
>
>       mcimage: mark_frame_dirty: failed to get lock
>
>(this happens before SKED starts.)

I think that this means that the McIDAS session is unable to create a
lock file in /tmp or was unable to write to ~user/.mctmp.

I logged onto your machine and became you and verified that the problem
was that the .mctmp directory (created by/for McIDAS) was not writable:

cd ~mkeables
ls -al
 ...
dr-x------   2 mkeables staff        512 Sep 21 14:06 .mctmp
 ...

I suspected that the cause for this was that the umask for your environment
was set to not create directories with write permission.  I tested this
by:

umask 002
mcidas

The McIDAS session then came up with no errors and .mctmp was created with
write permission:

> ls -al
 ...
drwx------   3 mkeables staff        512 Sep 30 14:55 .mctmp
 ...

It looks to me like you need to change the umask in your .cshrc file to
002.

>When I enter the following 
>command:
>
>       IMGDISP RTIMAGES/GEW-IR 1 
>
>I get the following response:
>
>       IMGDISP: Frame 1 does not exist
>       IMGDISP: Failed to write nav for frame 1
>       IMGDISP: loaded frame 1
>       IMGDISP: done
>
>So far, so good. When I try to draw a map on the image using
>
>       MAP H
>
>I get
>
>       MAP: NO IMAGE OR NO NAVIGATION FOR FRAME = 1

These errors are all caused from .mctmp not being writable.  After
setting umask and then running McIDAS, .mctmp was changed correctly
and I could run:

IMGDISP RTIMAGES/GEW-IR 1
Beginning Image Data transfer, bytes= 308208
IMGDISP: loaded frame  1
IMGDISP: done

>I don't get this problem when loging in as mcidas

That is due to the online instructions saying that umask should be set to
002.

>(nor do I get the error
>message about the frame at the start.

I did!?

>Suggestions?

Change umask to 002.

>Thanks in advance.

Later...

Tom

Date: Thu, 30 Sep 1999 08:13:03 -0600
From: Michael Keables <address@hidden>
Subject: Solaris NFS Problem

Hi.

We're trying to get cyclone.natnet.du.edu to function as a server for
McIDAS data products but can't seem to get cyclone to share the files.
Attached is a more detailed description of the problem by one of our
system administrators.

Suggestions?

Thanks in advance.

-- MK

From address@hidden  Tue Oct  5 13:08:17 1999

Based on the information below, I think issuing the command;

/etc/init.d/nfs.server start

or a system reboot will start NFS.  Afterwards, issuing the command;

showmount -e

will confirm that the partitions you selected are exported to the
system you listed (confirms no syntax errors).  Let me know if you
need any help or information.

mike

> >To: address@hidden
> >From: Michael Keables <address@hidden>
> >Subject: Solaris NFS Problem
> >Organization: University of denver
> >Keywords: 199909301417.IAA19834
>
> Hi.
>
> We're trying to get cyclone.natnet.du.edu to function as a server for
> McIDAS data products but can't seem to get cyclone to share the files.
> Attached is a more detailed description of the problem by one of our
> system administrators.
>
> Suggestions?
>
> Thanks in advance.
>
> -- MK
> Date: Wed, 29 Sep 1999 17:39:40 -0600 (MDT)
> From: Carl Gibbons <address@hidden>
> Subject: NFS problem (fwd)
>
> Hi!  I'll let you know when I get a reply... - Carl
>
> ---------- Forwarded message ----------
> Date: Wed, 29 Sep 1999 17:39:11 -0600 (MDT)
> From: Carl Gibbons <address@hidden>
> To: address@hidden
> Subject: NFS problem
>
>
>     Hi!  Mind if I ask another question?
>
>     A professor has an Ultra 10 box running Solaris 7 (cyclone), and a
>     couple of DELL-installed Red Hat Linux boxes (cumulus and nimbus).  We
>     want the Linux machines to NFS mount a directory that lives on the
>     Solaris machine.  When I try a mount, the Linux box coughs back:
>
> mount: RPC: Program not registered
>
>     I tried adding littlebear.cs.du.edu to the client list in the share
>     command in cyclone's dfstab, followed by a shareall.  Trying a mount
>     on littlebear (Solaris 2.6) gave me more detailed error information:
>
> nfs mount: cyclone.natnet.du.edu: : RPC: Program not registered
> nfs mount: retrying: /mnt
>
>     So I think the problem lies with the Solaris 7 box (cyclone),
>     not with the linux hosts.  I read some manpages and figured out
>     that I could use rpcinfo -p as a diagnostic tool.  Sure enough,
>     "rpcinfo -p | grep nfs" on cyclone returns no output, while on
>     a working NFS server (littlebear) I get lots of lines that look
>     similar to
>
> 100003    2   udp   2049  nfs
>
>     So I guess I need to do someting to rpc on the Solaris 7 host to make
>     it a well-behaved NFS server, but what?  Can you help?
>
>
> ------- End of Forwarded Message
>
>-- End of excerpt from Unidata Support


>From address@hidden  Mon Oct  4 17:14:29 1999

Hi, Tom.

Just wanted to let you know that we're finally up and running here at
DU. I've solved the NFS problem and now cumulus can access the data
files stored on cyclone.

Thanks a million for all of your help and patience with me during the
installation process!!

-- MK