This section covers some of the most commonly encountered errors when running the insertion scripts. They are divided into 3 separate tables, with each table handling errors originating from the LORIS-MRI installation (Table 1), the LORIS modules setup (Table 2), or the LORIS-MRI scripts (Table 3).
Table 1: Common errors encountered during LORIS-MRI installation, and their proposed solutions.
| Error | Cause | How to Fix |
|---|---|---|
install_driver(mysql) failed: Can't locate DBD/mysql.pm |
Missing dependency | sudo apt-get install libdbd-mysql-perl |
ERROR: You don't have a configuration file named 'prod' in: /data/%PROJECT%/bin/mri/dicom-archive/.loris_mri/ |
Your environment file does not contain your actual LORIS-MRI project name. Instead, it contains the placeholder %PROJECT% as provided in the 'generic' file and/or your environment file is not sourced |
Source the environment file located in /data/$PROJECT/bin/mri/ after making sure that the $PROJECT variable is replaced with your LORIS-MRI project name |
ERROR: You don't have a configuration file named 'prod' in: /data/loris-MRI/bin/mri/dicom-archive/.loris_mri/ note: loris-MRI is an example project name used in this illustration |
Wrong file and/or directory permissions | Make sure that the /data/$PROJECT/bin/mri directory, and all directories within are readable by the user running the scripts (lorisadmin or the front-end apache user) |
ERROR: You don't have a configuration file named 'prod' in: /data/loris-MRI/bin/mri/dicom-archive/.loris_mri/ note: loris-MRI is an example project name used in this illustration |
Syntax error in the prod file in the customized routines (for example a missing closing bracket) |
Check the routines that were customized for your project needs |
DB connection failed |
Database credentials in the prod file were entered incorrectly during the install, or they were modified subsequently |
Make sure that your prod file contains the correct database connection/credentials information in the DATABASE Settings, Section I |
Table 2: Common errors encountered due to missing LORIS (front-end) module setup steps, and their proposed solutions.
| Error | Cause | How to Fix |
|---|---|---|
| Images thumbnails do not show up in Imaging Browser. They appear as a broken image icon | Wrong permissions to the /data/$PROJECT/data/pic/ folder |
Ensure that the apache user can read/execute the pic images folder |
| Images thumbnails do not show up in Imaging Browser. They appear as a broken image icon | Wrong Images path under the Paths section in LORIS Configuration module |
Ensure the path to the images is correct, typically /data/$PROJECT/data/ |
| 4-D images (e.g. DTI, fMRI) in brainbrowser do not show any volumes (Play button not displayed) | Most likely a dcm2mnc conversion error | Post an issue on the minc-toolkit Github Issues page |
Brainbrowser says Loading… but no image shows up |
Wrong permissions to the /data/$PROJECT/data/assembly/ folder |
Ensure that the apache user can read/execute the MINC assembly images folder |
Brainbrowser says Loading… but no image shows up |
Wrong Images path under the Paths section in LORIS Configuration module |
Ensure the path to the MINC images is correct, typically /data/$PROJECT/data/ |
Brainbrowser says Loading… but no image shows up |
The config.xml in LORIS does not have the MINC Toolkit Path set properly |
Fill out the path <MINCToolsPath> to the MINC Toolkit Installation in the config.xml (on the LORIS side). The last trailing / in the path is mandatory |
Table 3: Common errors encountered during execution of the LORIS-MRi insertion scripts, and their proposed solutions.
| Error | Cause | How to Fix |
|---|---|---|
The Candidate info validation has failed |
PatientName/PatientID header in the DICOMs not anonymized according to the LORIS convention (PSCID_CandID_VisitLabel) |
Use DICAT to anonymize it properly OR Use the DICOM toolkit dcmodify command. The following one-line command (to be run from the folder where the DICOM files are) which anonymizes your entire folder of DICOM files is: for i in $(find -type f); do dcmodify -ma PatientName="PSCID_CandID_VisitLabel" -nb $i; done |
The Candidate info validation has failed |
The upload scan contains at least one file that is NOT of type DICOM (.bmp or .pdf are common) | Remove any file in the upload that is not of type DICOM |
... error message = 'No space left on device ...' |
The temporary directory where the insertion scripts perform its intermediate steps is full. This directory is set to a default value of /tmp as specified in the line export TMPDIR=/tmp of the environment file |
Change the TMPDIR path in the environment file to a directory with enough space. A good rule of thumb is to have at least 2-3 times the size of the scan being processed as writable space. This space is usually automatically emptied by the pipeline upon a successful execution |
ERROR: This class is designed around the notion of a 'Study'. You can't use it with data from multiple studies. The following study UIDs were found: '1.3.12.2.1107.66060.300000004' '1.3.12.2.1107.66060.300000007' The dicomTar execution has failed |
The upload contains acquisitions from two scanning sessions | Separate into two separate uploads |
Out of memory! during the execution of dicomTar.pl |
The Transfer syntax of the uploaded scan is other than Little Endian Explicit (such as Big endian, Little Endian Implicit or JPEG lossless) | First run dcmdump on one of the DICOM files within your archive. Look for the value of the tag TransferSyntaxUID. If it is of the form JPEGLossless* you will have to convert all the DICOM files in your archive so that the transfer syntax becomes Little Endian Explicit. You can use dcmdjpeg to do so, with a command like dcmdjpeg file.dcm file.dcm for each file. For those files with a transfer syntax other than JPEGLossless*, use the DICOM toolkit dcmconv to convert. An example command that changes the transfer syntax to Little Endian Explicit on all DICOM files within a given folder is: for i in $(find -type f); do dcmconv --write-xfer-little $i $i; done |
Number of MINC files that will be considered for inserting into the database: 0. No data could be converted into valid MINC files. Your project's excluded series descriptions will not be considered! |
The upload contains only acquisitions that have SeriesDescription which are excluded by the project |
Projects can specify what acquisitions' SeriesDescription are not to be processed by default in LORIS |
My uploaded DICOM study contains 6 modalities but only 5 were inserted. The 6th modality was not converted to MINC, neither inserted into the MRI Violated Scans module. In addition, during insertion, the message Number of MINC files that will be considered for inserting into the database: 5 is displayed, confirming that only 5 files are considered |
The missing modality is probably matching one of the excluded series descriptions set in the Config Module under the Imaging Pipeline section. tarchiveLoader is specifically 'instructed' to exclude those modalities |
Make sure these excluded series descriptions have correctly been set (note, they need to be entered as an exact match to what is present in the DICOM file) |
My uploaded DICOM study contains 6 modalities but only 5 were inserted. The 6th modality was not converted to MINC, neither inserted into the MRI Violated Scans module. In addition, during insertion, the message Number of MINC files that will be considered for inserting into the database: 5 is displayed, confirming that only 5 files are considered |
Probably the DICOM headers have blank SeriesNumber. LORIS-MRI generates MINC files and names them temporarily (in intermediate insertion steps) based on their SeriesNumber, before it proceeds to renaming them once the protocol is identified from the mri_protocol table. Having no SeriesNumber defaults to 0, causing overwrites on the intermediate MINC files generated |
Ensure that your DICOM headers include a non-blank SeriesNumber for every acquisition in that specific study |
The target directory does not contain a single DICOM file |
Probably the DICOM headers have blank StudyUID. The logic of insertion within LORIS-MRI depends on a StudyUID header | Ensure that your DICOM headers include a non-blank StudyUID header |
My resting-state fMRI scans are tagged as task fMRI although I have 2 entries in the mri_protocol table |
The resting-state scan has parameters that match those of the task entry of the mri_protocol table, and the task-related entry in the mri_protocol table precedes that of the resting-state fMRI |
Make sure the mri_protocol table has parameters that discern between all the study acquired modalities in an exclusive manner (i.e. no two row entries have overlapping parameters across all their columns) |
no MINCs inserted |
Possibly all the MINC images are violated scans | Check the details of the image headers (from the MRI Violated Scans module or using mincheader) against the mri_protocol table entries, and adjust the table protocol parameters accordingly |
Key configuration points to verify:
-
Depending on your operating system, some dependencies might be missing. During initial troubleshooting of the imaging pipeline, note any related error messages (e.g.
install_driver(mysql) failed: Can't locate DBD/mysql .pm) and install missing packages as needed (e.g.sudo apt-get install libdbi-perl,sudo apt-get install libdbd-mysql-perl,sudo apt-get install libarchive-zip-perl). -
If your MINC toolkit is older than 1.9.14 and your scans have no Date of Birth value, you may see an age unit error during DICOM to MINC conversion. Instructions for compiling a more recent version of the MINC toolkit are available on MNI-BIC GitHub.
Once an MRI scan session has been successfully inserted, it will be listed in the Imaging Browser main data table (also linked from the DICOM Archive sub-page: "View Images").
Verify in the Imaging Browser's View Session page that a jpg showing 3 slice
orientations displays properly; if not, verify your permissions and restart
apache:
sudo chmod o+r /data/$PROJECT/bin
sudo chmod o+r /data/$PROJECT/data
sudo service apache2 restart
If download links do not work, ensure that the /data/$PROJECT/data/assembly
directory and subdirectories are executable.
Error and output messages from the imaging insertion scripts are logged in files
created under the /data/$PROJECT/data/logs/ directory. To view messages from
the last script run, consult the most recent log file modified in this
directory. These log files reference an uploadID used to identify each
imaging dataset -- consult the mri_upload database table to look up which
uploadID has been assigned to your scans.
Caveat: When the imaging insertion pipeline is auto-launched by the
Imaging Uploader module, the pipeline scripts' log files are output to
/tmp/ and deleted. To avoid deletion, edit the Server Processes Manager
function deleteProcessFiles()
to return false instead of true.
If upload was successful but issues were encountered with the imaging insertion pipeline scripts:
- CentOS: check for additional dependencies/configurations (e.g. DICOM Dictionary path) in the detailed CentOS Imaging Installation transcript
- Manually re-run the entire pipeline sequence using the
imaging_upload_file.plscript - If one of the final steps such as the MINC conversion is failing, you may
wish to just re-run the
tarchiveLoaderscript. - See also re-running the Imaging pipeline section for troubleshooting information.