- galaxy-dev - lists.galaxyproject.org

Deleted jobs on the cloud keep running
by Jorrit Boekel 29 Aug '12

29 Aug '12

Dear list, When running Galaxy/Cloudman on Amazon (on Ubuntu 12.04), I run into the following: - I start a job with <parallelism> in the tool.xml file. - Since I have use_tasked_jobs=True in universe_wsgi.ini, the job is split and divided over the instances on EC2 - I change my mind, and click on the delete button in the Galaxy UI. This removes the job from the UI. - I log in to an instance running my job, and see that it is still running (both its wrapper and the tool program). It keeps running, keeping my instances busy. - When tasks of this job finish (I assume), it is reported to galaxy, and my logfile says things like: galaxy.jobs DEBUG 2012-08-29 12:42:08,239 task 319 for job 25 ended galaxy.jobs DEBUG 2012-08-29 12:42:08,265 The tool did not define exit code or stdio handling; checking stderr for success galaxy.jobs DEBUG 2012-08-29 12:42:08,282 task 319 ended So I wonder if the default behaviour when deleting jobs in the UI is that they keep running but hidden, or is there something wrong in my config? Is there a way to change this with a setting? cheers, jorrit boekel

2 1

Running apache as galaxy
by Joachim Jacob 29 Aug '12

29 Aug '12

Dear all, I had problems configuring apache, which were solved by running apache as user galaxy. Is this an security issue? The server will only be use for galaxy. Cheers, Joachim -- Joachim Jacob, PhD Rijvisschestraat 120, 9052 Zwijnaarde Tel: +32 9 244.66.34 Bioinformatics Training and Services (BITS) http://www.bits.vib.be @bitsatvib

1 0

How to restart the Galaxy process
by kauerbach＠comcast.net 29 Aug '12

29 Aug '12

Hello, I modified the Galaxy configuration file and would like to stop and re-start the Galaxy process. Can you tell me how to do that? Thank you.

2 1

Using the tools API
by Clare Sloggett 29 Aug '12

29 Aug '12

Hi guys, The Tools API is currently working for me from galaxy-central, but I'm not sure how to correctly run a tool. Are there any example scripts, as there are for some other parts of the API? Specifically I want to find out what the expected payload fields are when I post to CREATE to run a tool. Some of the fields are clear to me just from the api/tools.py code (e.g. 'tool_id') but others are not (e.g. how the input datasets and parameters are specified). A separate question: How do we specify "Advanced" or conditional-dependent fields for a tool? Some of these fields are necessary to run the tool at all. For instance, on my system, calling http://localhost:8080/api/tools/tophat?key=.... returns { "description": "Find splice junctions using RNA-seq data", "id": "tophat", "inputs": [ { "html": "%3Cselect%20name%3D%22input1%22%3E%0A%3C/select%3E", "label": "RNA-Seq FASTQ file", "name": "input1", "type": "data" }, { "label": "Conditional (refGenomeSource)", "name": "refGenomeSource" }, { "label": "Conditional (singlePaired)", "name": "singlePaired" } ], "name": "Tophat for Illumina", "version": "1.5.0" } This is obviously only some of the inputs you see in the UI. I think that all the Advanced fields are missing, and more importantly, any input which is dependent on a conditional is missing. So the refGenomeSource conditional is there, but the actual reference genome field is not. The type of the reference genome field also presumably depends on which value is supplied for the referenceGenomeSource conditional. Is there currently a way to specify (or see) these missing fields? Thanks, Clare -- Clare Sloggett Research Fellow / Bioinformatician Life Sciences Computation Centre Victorian Life Sciences Computation Initiative University of Melbourne, Parkville Campus 187 Grattan Street, Carlton, Melbourne Victoria 3010, Australia Ph: 03 903 53357 M: 0414 854 759

2 4

Pull request, and missing step connections
by Clare Sloggett 29 Aug '12

29 Aug '12

Hi guys, I made a very small code change so that the workflows API will display all the steps in a workflow, not just the inputs (these are still displayed separately, as before). I made a pull request even though the change is small, just to learn what I'm doing: https://bitbucket.org/galaxy/galaxy-central/pull-request/68/show-workflow-s… But in doing so I noticed something in the workflows I don't understand. I can't see a connection between workflow "input datasets" and the steps they are inputting to. I doubt this is a bug, I just don't know how it's supposed to work, so I'm not sure if my API change is sufficient. So for instance, if I create a small workflow with steps Input Dataset -> TopHat (accepted_hits bam file) -> Cufflinks and call the API on this workflow, I now see { "id": "f2db41e1fa331b3e", "inputs": { "1": { "label": "Input SE fastq", "value": "" } }, "name": "Tophat + cufflinks", "steps": { "1": { "id": 1, "input_steps": {}, "tool_id": null, "type": "data_input" }, "2": { "id": 2, "input_steps": {}, "tool_id": "tophat", "type": "tool" }, "3": { "id": 3, "input_steps": { "input": { "source_step": 2, "step_output": "accepted_hits" } }, "tool_id": "cufflinks", "type": "tool" } }, "url": "/api/workflows/f2db41e1fa331b3e" } The "inputs" field was there before, the "steps" field is the new bit. So as expected, step 3 lists step 2 as an input. However step 2 does not list step 1 as an input, even though the GUI shows that they are connected and the workflow works. From other testing it seems that "Input Dataset" steps never appear wired up in my API response. I'm just iterating over all input_connections, so apparently steps of type data_input are not in the list of WorkflowStep.input_connections . Is this how it should be? How can I find out which steps an Input Dataset is connected to? Thanks, Clare -- Clare Sloggett Research Fellow / Bioinformatician Life Sciences Computation Centre Victorian Life Sciences Computation Initiative University of Melbourne, Parkville Campus 187 Grattan Street, Carlton, Melbourne Victoria 3010, Australia Ph: 03 903 53357 M: 0414 854 759

1 0

DataTypes
by Alfredo Guilherme Silva Souza 28 Aug '12

28 Aug '12

Good Afternoon. I need to create a new data type in the galaxy, because I need to import a file in the format. Pdb. Someone already had to create a new kind of data? Can you help? Thank you. -- *Alfredo Guilherme* * *

2 1

problems with labels in the tool box and order of the tools
by Hans-Rudolf Hotz 28 Aug '12

28 Aug '12

HI We have been struggling with this problem since we updated to "changeset:6799:40f1816d6857" (March 12, 2012 release). We never had the time to figure out whether it was a bug or whether we made a mistake. Unfortunately, it now has reached a point where we have a serious problem with our production server. I managed to reproduce the error on a new galaxy-dist clone "changeset:7404:ec29ce8e27a1" (July 20, 2012 release): I modify the top of "tool_conf.xml" in the following way (ie: adding a label): <?xml version="1.0"?> <toolbox> <section name="Get Data" id="getext"> <tool file="data_source/upload.xml"/> <label text="external" id="external_data" /> <tool file="data_source/ucsc_tablebrowser.xml" /> <tool file="data_source/ucsc_tablebrowser_test.xml" /> ... and it works, after restart, I get the label (see attachment "ss2.tiff") I add a new tool definition: "load_abi_reads.xml": <?xml version="1.0"?> <toolbox> <section name="Get Data" id="getext"> <tool file="data_source/upload.xml"/> <tool file="hrh/load_abi_reads.xml"/> <label text="external" id="external_data" /> <tool file="data_source/ucsc_tablebrowser.xml" /> <tool file="data_source/ucsc_tablebrowser_test.xml" /> ... and it works, after restart, I get the new tool (see attachment "ss3.tiff") Now, I don't like the label anymore <toolbox> <section name="Get Data" id="getext"> <tool file="data_source/upload.xml"/> <tool file="hrh/load_abi_reads.xml"/>  <tool file="data_source/ucsc_tablebrowser.xml" /> <tool file="data_source/ucsc_tablebrowser_test.xml" /> but, after restart, the label is still there. Even worse, I add a second tool ("load_novaseq_reads.xml") before the first added tool <?xml version="1.0"?> <toolbox> <section name="Get Data" id="getext"> <tool file="data_source/upload.xml"/> <tool file="hrh/load_novaseq_reads.xml"/> <tool file="hrh/load_abi_reads.xml"/>  <tool file="data_source/ucsc_tablebrowser.xml" /> <tool file="data_source/ucsc_tablebrowser_test.xml" /> and after restart, the second tool is put below the first added tool (and the label is also still there, see attachment "ss4.tiff") The display in the web browser corresponds to the file "integrated_tool_panel.xml", eg: <?xml version="1.0"?> <toolbox> <section id="getext" name="Get Data" version=""> <tool id="upload1" /> <tool id="load_abi_reads" /> <tool id="load_novaseq_reads" /> <label id="external_data" text="external" version="" /> <tool id="ucsc_table_direct1" /> <tool id="ucsc_table_direct_test1" /> <tool id="ucsc_table_direct_archaea1" /> What do I need to change, in order to have the same information in the "integrated_tool_panel.xml" file as in the "tool_conf.xml" file with regard to order of the tools and removal of labels? Any help appreciated Thank you very much Regards, Hans -- Hans-Rudolf Hotz, PhD Bioinformatics Support Friedrich Miescher Institute for Biomedical Research Maulbeerstrasse 66 4058 Basel/Switzerland

3 7

Re-starting Galaxy
by kauerbach＠comcast.net 28 Aug '12

28 Aug '12

Hello, If I were to stop and re-start the Galaxy server with the run.sh script would jobs that are aleady running or are in the queue be lost when the server is re-started? If so, is there another method to use which would not lose the running or queued jobs? Thank you.

5 4

ToolShed enhancement for repository lists
by Peter Cock 28 Aug '12

28 Aug '12

Hi all, This is an idea I mentioned to Dannon (off list during discussion about moving the NCBI BLAST+ wrappers to the Tool Shed), and he seemed positive about it. Currently the left hand pane of the ToolShed (once logged in) looks like this: Search * Search for valid tools * Search for workflows Repositories * Browse by category * Browse my repositories * Browse my invalid tools * Create new repository I would find it useful to add another entry listing those repositories where I have contributor rights. This could be included under the existing "Browse my repositories", but perhaps a separate entry is better - if anyone has a concise description for this (so far I've not come up with a short way of phrasing this). Should I file this as an enhancement issue on BitBucket? Regards, Peter

2 1

potential operate on genomic intervals-intersect bug
by Michael Wilson 28 Aug '12

28 Aug '12

Dear Galaxy, Last night (1AM) and today (10:45AM) I have not been able to return results with the operate on genomic intervals, intersect tool. For example if I upload ChIP-seq peak replicates in .gz format (downloaded directly from ENCODE), no results are returned when they are intersected with each other (or with themselves). If I use the Join tool I indeed am able to join many intervals as would be expected. Not sure if I am doing something wrong this time..... Here is a shared history link to give an example: https://main.g2.bx.psu.edu/u/mdwilson/h/intersect-test Best regards, Mike

1 0