see title

I'm having some trouble submitting a matlab job to the queue. The command I use is:

/software/common/bin/ezsub -c "matlab -nodesktop -nosplash < /groups/vwfa/att_pilot/batch/PPI/runPPIFTWComplex.m" -n ftwrun

If I simply run matlab from command line: matlab -nodesktop -nosplash < /groups/vwfa/att_pilot/batch/PPI/runPPIFTWComplex.m

it runs fine, but whenever I submit it to the queue, it fails right away with exit status=139.

when I open up a terminal, I get the error:

bash: fss: No such file or directory

I think this is in response to these lines in my .bashrc file

source /etc/fsl/fsl.sh
. fss 5.0.0

Is this a problem?

How would we run:

recon-all -base template_SAD_019 -tp SAD-019 -tp SAD2_019 -all

on the cluster? We tried:

recon-all-pbs -base template_SAD_019 -tp SAD-019 -tp SAD2_019 -all -q recon

but it didn't work!

does load balancing on the recon queue take memory consumption into account? or is it purely based on cpu usage? if cpu only, could we create a bigmem queue that takes jobs which use upto 10GB of RAM and will only run at most 4 of them on each of the mh nodes.

need help getting access to om/project/sart

adamjhh@openmind7

Hi! I have an issue:
When I try to open a node the systems asks for a password. It does not accept my password, but if I simply press return 3 times it looks like it gives me access to the node. However, an error message says that x11 could not connect to the node:

[17:51][1.85][-102%]xaviergp@openmind7:~$ srun --x11 --pty --mem=1G bash
The authenticity of host 'node012 (172.16.20.12)' can't be established.
ECDSA key fingerprint is b8:77:68:03:fc:8d:a4:5a:64:87:7a:9a:10:10:9a:90.
Are you sure you want to continue connecting (yes/no)? 
xaviergp@node012's password: 
xaviergp@node012's password: 
xaviergp@node012's password: 
srun: error: x11: unable to connect node node012
[17:52][14.31][-88%]xaviergp@node012:~$

Once I am in the node I cannot open windows such as Matlab. It seems like the reason is that x11 is not working well in the node:

[17:52][14.31][-88%]xaviergp@node012:~$ module add mit/matlab/2016b
[17:52][14.52][-88%]xaviergp@node012:~$ matlab
MATLAB is selecting SOFTWARE OPENGL rendering.

                                       < M A T L A B (R) >
                             Copyright 1984-2016 The MathWorks, Inc.
                              R2016b (9.1.0.441655) 64-bit (glnxa64)
                                        September 7, 2016

 
To get started, type one of these: helpwin, helpdesk, or demo.
For product information, visit www.mathworks.com.

Is there any solution?
Thank you!
Xavier.

when submitting a job to execute a python script that uses matplotlib to create figures, I get the error "Unable to access the X Display, is $DISPLAY set properly?".
I use NX and submitted the job with qsub -V.
$DISPLAY was first 50002.0 but setting "export DISPLAY=ba3.mit.edu:50002.0" did not solve anything.

I am running matlab on ba3 and could not use any functions in the stats toolbox.
Got these warning messages for toolboxes when starting up matlab.
....
/software/matlab_versions/2010b/toolbox/distcomp/parallel/datafun.
Warning: Name is nonexistent or not a directory:
...

it's unclear why i would have lost credential.

INFO:interface:stderr 2011-10-14T20:06:19.923592:qsub: Invalid credential
Traceback (most recent call last):
  File "resting_rois.py", line 130, in <module>
    roiproc.run('PBS', plugin_args={'qsub_args' : '-o /dev/null -e /dev/null'})
  File "/software/python/EPD/virtualenvs/7.1/nipype0.4/lib/python2.7/site-packages/nipype/pipeline/engine.py", line 468, in run
    runner.run(execgraph, updatehash=updatehash, config=self.config)
  File "/software/python/EPD/virtualenvs/7.1/nipype0.4/lib/python2.7/site-packages/nipype/pipeline/plugins/base.py", line 159, in run
    self._send_procs_to_workers(updatehash=updatehash)
  File "/software/python/EPD/virtualenvs/7.1/nipype0.4/lib/python2.7/site-packages/nipype/pipeline/plugins/base.py", line 243, in _send_procs_to_workers
    updatehash=updatehash)
  File "/software/python/EPD/virtualenvs/7.1/nipype0.4/lib/python2.7/site-packages/nipype/pipeline/plugins/base.py", line 405, in _submit_job
    return self._submit_batchtask(batchscriptfile, node)
  File "/software/python/EPD/virtualenvs/7.1/nipype0.4/lib/python2.7/site-packages/nipype/pipeline/plugins/pbs.py", line 50, in _submit_batchtask
    str(e))))
RuntimeError: Could not submit pbs task for node register.a520
Command:
qsub -o /dev/null -e /dev/null -N satra.register.a520 /g2/users/satra/ADHD200/working/adhd200rois/batch/batchscript_20111014_200619_register.a520.sh
Standard output:

Standard error:
qsub: Invalid credential
Return code: 173
Interface CommandLine failed to run. 

(see title)

Hey guys,
This is not really an issue, so I'm not sure if this is the best place to ask this... Should I post to the gablab list instead? the nipype mailing list?
I've never used rapid art as my artifact detection tool before and I was hoping to find more information than the following about how it works, including the math behind it and perhaps also a notion about what good thresholds to use are (especially in different populations).

Thanks! Sorry if this is the wrong place for this question!!

Amy

http://nipy.sourceforge.net/nipype/api/generated/nipype.algorithms.rapidart.html#module-nipype.algorithms.rapidart

It seems that Matlab can't be started when I'm logged onto ba3

the new test website http://gablab.mit.edu/joomla_test/index.php/en/ doesn't seem to accept my LDAP user/password. Is this an issue to bring up with Matt?

I submitted a pair of jobs last night which don't seem to be actually running, but which still appear in qstat. In particular, 54218.sub and 54219.sub They are not removed by trying 'qdel' or 'qsig' (though neither command returns an error).

Two questions: 1) How to get such dead jobs out of the qstat, and 2) what is the 'best practices' way to kill a running job in the queue if you want it to stop?

Hi all,
looks like there are no free display numbers for vnc sessions. I noticed b/c I wanted to kill mine and start a new one as I can't even open a terminal window right now. Any idea what is going on. NX is an ok back up for now.

when running df -h /mindhive/gablab it says there is NO space left anymore, 0! and i can't even create an empty text file anymore. i tried freeing some space but nothing changed after i cleared like 4G (i know its not much, but it should be enough to let me do "mv tinyfilea tinyfileb"?)

When I log into ba3 using ssh -X from a mac, I can't open freeview. People are able to open freeview when ssh'ing in from a linux system, or using vnc.

from iTerm on a mac:

ssh -X ba3.mit.edu
(initialize freesurfer stuff)

$ which freeview
/software/Freesurfer/5.0.0/bin/freeview

$ freeview
Xlib: extension "RANDR" missing on display "localhost:18.0".
The program 'freeview.bin' received an X Window System error.
This probably reflects a bug in the program.
The error was 'BadValue (integer parameter out of range for operation)'.
(Details: serial 19 error_code 2 request_code 146 minor_code 3)
(Note to programmers: normally, X errors are reported asynchronously;
that is, you will receive the error a while after causing it.
To debug your program, run it with the --sync command line
option to change this behavior. You can then get a meaningful
backtrace from your debugger if you break on the gdk_x_error() function.)

(gvim:11267): GnomeUI-WARNING *: While connecting to session manager:
networkIdsList argument is NULL.
*
GLib-GIO:ERROR:/build/buildd/glib2.0-2.28.6/./gio/gdbusconnection.c:2279:initable_init: assertion failed: (connection->initialization_error == NULL)
Vim: Caught deadly signal ABRT
Vim: Finished.
Aborted

The cluster seems to be barely trickling jobs through, keeping them in the queued state. Also and possibly related, pbsnodes is outputting strange unicode characters in its list of jobs.

(nipype0.4)satra@ba3:/software$ mkdir mindhive
mkdir: cannot create directory `mindhive': Permission denied

Following the upgrade to ba3, I am now getting a number of errors when I ask for forwarded X-packets to get GUI applications over SSH. See the terminal output below. The forwarded windows do not always render properly (for example, in gedit, the tabs for different open files are all black, rather than showing text on the system background color). Other applications are affected too.

[13:16][8.56]tkp@ba3:~$ gedit /.bashrc &
[1] 26985
[13:17][8.52]tkp@ba3:$ Gtk-Message: Failed to load module "pk-gtk-module"

** (gedit:26985): WARNING **: Could not load theme icon gtk-home: Icon 'gtk-home' not present in theme

(gedit:26985): LibZeitgeist-CRITICAL **: Failed to create proxy for Zeitgeist daemon: Error calling StartServiceByName for org.gnome.zeitgeist.Engine: GDBus.Error:org.freedesktop.DBus.Error.Spawn.ChildExited: Process /usr/bin/zeitgeist-daemon exited with status 1

(gedit:26985): LibZeitgeist-CRITICAL **: Unable to connect to Zeitgeist daemon: Error calling StartServiceByName for org.gnome.zeitgeist.Engine: GDBus.Error:org.freedesktop.DBus.Error.Spawn.ChildExited: Process /usr/bin/zeitgeist-daemon exited with status 1

(gedit:26985): GLib-GIO-CRITICAL **: g_dbus_proxy_get_connection: assertion `G_IS_DBUS_PROXY (proxy)' failed

(gedit:26985): GLib-GObject-CRITICAL **: g_object_ref: assertion `G_IS_OBJECT (object)' failed

(gedit:26985): GLib-GIO-CRITICAL **: g_dbus_connection_signal_subscribe: assertion `G_IS_DBUS_CONNECTION (connection)' failed

Not a pressing issue, but a bit annoying.

Many of the man pages for common shell commands and programs aren't found.

tkp@ba3:~$ man ls
No manual entry for ls

tkp@ba3:~$ echo $MANPATH
/software/pkg/torque/torque-2.5.8/man:/usr/man
tkp@ba3:~$ ls /usr/man
ls: cannot access /usr/man: No such file or directory

(in a way it's rather poetic -- man can't find ls, and ls can't find man).

(It's an spm batch script)

df says /scratch is 97% full - in fact it was 100% full till i removed some data.

stm:/scratch 9.9T 9.1T 306G 97% /mindhive/scratch

while du says:

(nipype0.4)satra@ba3:/mindhive/scratch$ du -hs
567G .

qstat
pbs_iff: cannot read reply from pbs_server
No Permission.

When I log in to Openmind through X2Go I just see a blank screen. Can you reboot my Centos?

Separately, when I call fsl I get an error 'Cannot connect to X server'

Thank you

Sorry if this is documented elsewhere. Does NX work to get onto ba3 any longer? Its failing for me, but I may have missed details regarding how to change my settings. Also - in another problem posted, a work around for X was given. Does this mean we can no longer NX in?
A

In matlab2010b, the stats toolbox isn't loading. This wasn't a problem in the past (with 2010b), so it's definitely installed, but it's probably not loading due to licensing. Perhaps there are fewer licenses than users? It isn't listed in ver and isn't present in the toolbox list.

cannot cd into /mindhive/gablab/sad for example.

I'm connected to ba3 now via VNC and looking in /mindhive/gablab, and all I see are the directories from different partitions, not the actual /mindhive/gablab/ partition (which if I remember correctly is munin:/vol/vol1, but don't quote me on that).

tkp@ba3:gablab$ ls
fluid  memory  pbc  reading  rtsmoking  sad  zeynep

Are the machines running slower than usual? I'm running a pypeline script in interactive mode on ba9 and it's taking on the order of minutes just to scroll past the screen output before the nodes actually start running. I also have some other scripts in the queue that have been running estimation nodes for 27 hours now and still going.

Three jobs running on the 'recon' queue are running on ba6. I was under the impression this queue would direct its jobs to the mh nodes?
Jobs are 542206, 542207, 542208

Hi everyone,

I'm using my laptop at home and trying to look at some volumes but freeview gives me this unhelpful error:

(nipype0.4)jsalva@ba3:/mindhive/gablab/users/jsalva$ which freeview
/software/Freesurfer/5.1.0/bin/freeview
(nipype0.4)jsalva@ba3:/mindhive/gablab/users/jsalva$ freeview
X Error: BadValue (integer parameter out of range for operation) 2
Extension: 147 (Uknown extension)
Minor opcode: 3 (Unknown request)
Resource id: 0x39
Abort

Does anyone have a suggestion/solution? Thanks

John

I'm logged into ba3, but can't start gvim. Apparently it's not installed.