kellyuw / vmpfc-notes Goto Github PK

rPkgsInstall -pkgs ALL -check

oo Warning:
These packages are not installed on the computer: afex!
These packages are not installed on the computer: phia!
These packages are not installed on the computer: snow!
These packages are not installed on the computer: lme4!
These packages are not installed on the computer: paran!
These packages are not installed on the computer: psych!

AFNI needs to be updated to the new version that is currently installed on other cluster machines (16.0.11). This will help to ensure that processing is consistent across various IBIC SGE nodes and also allow us to take advantage of important recent bug fixes.

Unable to view images in ITK-SNAP via X2Go.

Very soon after the extra machines were unmounted from vmpfc (7/07/16 in reference to issue #2 ), I noticed that I was unable to open a new terminal on my desktop workstation (the normal terminal window display appeared, but there was no prompt available for entering text/commands). SSH from vmpfc to other workstations and cluster machines also seemed to be unsuccessful.

Some scripts (e.g. subject_setup.sh) rely on specific version of MATLAB for conversion with r2aGUI. New cluster machines have newer version of MATLAB, so we will plan to upgrade MATLAB on vmpfc to match new version on cluster machines (2016a).

Jobs stuck in queue for very long periods of time (24+ hours) and do not appear to be making any progress. New jobs remain stuck in pending queue.

Rstudio is nearly impossible to find if not included in path!

strace is a helpful debugging utility that should be installed on vmpfc (to be consistent with other cluster machines).

Not able to access home directory (since ~10:30am today). This has happened several times in the last week or so and often kills jobs (particularly those that require use of isolated python environment via anaconda).

Originally noticed 6/29/2016. New SGE jobs (submitted via submit_parallel_pipeline script and/or parallelized FSL jobs) stuck in pending queue (seemingly forever).

With the intended automounter configuration, the mount of a machine to /mnt/ will only occur when it is in use. There is a timeout that will specify when it should be unmounted after a period of disuse. “Use” can mean cd’ing into a directory and staying there, or viewing an image within the mounted directory with fslview (specifically, “in use” means that a file is open under that mountpoint).

The puzzling thing about vmpfc is that all neuron machines were mounted, including a few that we didn't know of previously (chdd, neuroimaging2). This is highly unexpected because no one should be using those volumes on vmpfc. Panuc was mounted on vmpfc but panuc was not correctly exporting, and so caused the load to spike (artificially - this happens with our NAS too) and thereby trigger an alarm on the SGE that prevented jobs from running on vmpfc ( #1 )

All ssh connections to vmpfc terminated on Wednesday (7/20/2016) around 12:30pm. Not able to reconnect via ssh or x2go.

Received 'Destination Host Unreachable' error message in response to attempts to ping vmpfc from dws-12.
PING vmpfc.ibic.washington.edu (128.95.48.122) 56(84) bytes of data.
From dws-12.ibic.washington.edu (128.95.48.86) icmp_seq=1 Destination Host Unreachable

Inaccessible this morning. Heard report that it may be related to earlier issues with UPS. TPP was brought online again by 11:30am, but still waiting on vmpfc...