January 2017 - galaxy-dev - lists.galaxyproject.org

Question about Data Managers
by Carlos Lijeron 03 Feb '17

03 Feb '17

Everyone, I’m trying to setup various data managers to automatically download and index reference indices. However, I can’t figure out where to change the default location where these are saved. As you can see in the attached image, the default location for the indexes is the … /tool-data/dbkey/index_name However, I would like to save them in /projects/reference_datasets which is a shared location that has a lot of space available. Please advise. Thank you all ! Carlos. [cid:image001.png@01D25B9D.4F483110]

2 1

Check running jobs
by Katherine Beaulieu 30 Jan '17

30 Jan '17

Hi there, We are trying to build our own Galaxy cluster and I was wondering if there was a way to check if there are any running Galaxy jobs on a specific server? This would be to determine whether or not to remove the server from the cluster. Thanks! Katherine

2 4

Problem with 'galaxy_packages' dependency resolver in versionless mode
by Peter Briggs 26 Jan '17

26 Jan '17

Dear developers I'm having problems trying to use the 'galaxy_packages' dependency resolver in versionless mode, as described in the docs at: https://docs.galaxyproject.org/en/latest/admin/dependency_resolvers.html#ga… As I understand this: if I manually create a directory structure of the form e.g. /path/to/local_tool_dependencies/my_package/default/ which contains an env.sh file, and if my dependency_resolvers_conf.xml file includes the line: <galaxy_packages base_path="/path/to/local_tool_dependencies" versionless="true" /> then at run time Galaxy should be able to resolve the following requirement in a tool XML file: <requirement type="package">my_package</requirement> Unfortunately although the "versioned" galaxy_packages resolver works as expected (when placed ahead of the "versionless" resolver), I don't seem able to get this working with Galaxy v16.10. Can anyone point out what I'm doing wrong? Thanks! Best wishes Peter -- Peter Briggs peter.briggs(a)manchester.ac.uk Bioinformatics Core Facility University of Manchester B.1083 Michael Smith Bldg Tel: (0161) 2751482

3 3

impersonation doesn't work in v16.10
by Ping Luo 25 Jan '17

25 Jan '17

I recently installed galaxy v16.10 and have set "allow_user_impersonation = True". However, the impersonation feature seems don't work properly. When I click "Admin->impersonate a user" and choose the account I want to impersonate, and the click "impersonate" button. The page shows "you have login as xxx, return to the homepage". But when I click homepage, the actual login user is still my account. I have been using galaxy v1507 and I have never had any problem with impersonation. I am not sure if this is a bug in v1610 or I just miss something. Ping

1 0

How to satisfy a Galaxy tool dependency on python-irodsclient
by Lawrence, Cris W 24 Jan '17

24 Jan '17

I am a Galaxy noob trying to write a Galaxy tool that needs to employ the python-irodsclient module. I can install it on my laptop via pip. But I am trying to make it available to Galaxy via planemo. Apparently, this module is not available in the conda channels. How else can I satisfy my dependency? Thanks for any and all help. - Cris

3 7

Which "upstream" Ansible Playbook to use for local galaxy instance?
by Hendrickson, Curtis (Campus) 24 Jan '17

24 Jan '17

Dear Galatic Advisors, We're building a new set of machines to replace our current local galaxy instance (prod, dev and db VMs running CentOS against a local SLURM cluster with local storage, want to start using wsgi), and are planning to start using Ansible, instead of doing everything by hand. Looking through the wikis and postings, we see many difference points of departure. Can anyone outline the pros and cons of these various choices or provide any recommendations? Is it best to use an edited version of this: https://github.com/galaxyproject/usegalaxy-playbook Or this: https://github.com/galaxyproject/ansible-galaxy I also see a 3rd version that is linked from the "setting up your own galaxy" page (wiki) https://wiki.galaxyproject.org/Admin/GetGalaxy (Playbook) https://github.com/ARTbio/GalaxyKickStart Ideally, we'd like to benefit from updates to the upstream package. All three seem to be actively supported. Thanks Curtis @ UAB.edu PS - is it better to post this on galaxy-dev or on https://biostar.usegalaxy.org?

1 0

display at UCSC does not work
by Hans-Rudolf Hotz 24 Jan '17

24 Jan '17

Hi I am trying to set up the possibility to display data sets with the UCSC genome browser for our new galaxy installation. After clicking on "display at UCSC main", I get the UCSC genome browser, opening in the right assembly, and at the correct location, but I get the following Error: Warning/Error(s): redirected to non-http(s): /login?redirect=%2Froot%2Fdisplay_as In galaxy.ini, I did uncomment the line: display_servers = hgw1.cse.ucsc.edu,hgw2.cse.ucsc.edu,hgw3.cse. ... We use the 'new' authentication method to connect to our LDAP server, with the following changes in galaxy.ini: require_login = True auth_config_file = config/auth_conf.xml Connection from UCSC to our galaxy server is working, since I can display the data, when I switch to require_login = False it works, independent of whether I am logged in or not. Is there a way to get it to work, even if I have "require_login = True"? Thank you very much for your help Regards, Hans-Rudolf -- Hans-Rudolf Hotz, PhD Bioinformatics Support Friedrich Miescher Institute for Biomedical Research Maulbeerstrasse 66 4058 Basel/Switzerland

1 1

Resource temporarily unavailable
by Durdevic, Marija 24 Jan '17

24 Jan '17

Dear developers, >From time to time I have issue on my Galaxy server. The problem is when job is done and results has to be written to the database something happened. The error message is: "/home/data/galaxy_database/files/024/dataset_24790.dat: Resource temporarily unavailable" where "/home/data/galaxy_database/files/024/dataset_24790.dat" is path to the output file. Although it produced reasonable output the job is shown as failed. I am running galaxy over Slurm and we use NFS for connection with storage. Do anyone have idea what can be possible reason for such error? Thanks in advance. Best, Marija

2 2

Synnefo Storage Backend for Galaxy
by Stavros 0 Sachtouris 20 Jan '17

20 Jan '17

Hello, my name is Stavros Sachtouris and I'm developing an object storage backend for Synnefo [1] Object Store (aka Pithos+ [2]). In the context of the OpenMinTeD (Open MINing for TExt and Data) project [3], we are trying to use Galaxy as a workflow engine and we need to set it up on to use our Object Store (Pithos+). Therefore, we need a Pithos+ backend for Galaxy. Pithos+ API claims compatibility with OpenStack Swift API, but in practice this rarely works. That's why we have developed our own python API library (kamaki [5]). In order to provide Pithos+ access to Galaxy, I had to develop a Pithos+ backend using kamaki. We then deployed a testing Galaxy and patched it with our code and it seems to work well. I'm sending this message for two reasons: a. to let the Galaxy developers know about our interest in Galaxy and start a cooperation which is hopefully going to be beneficial for both parties. b. to ask for advice and guidelines on how to prepare my first patch and how to ship it for review. Although the patch seems to work, we feel like we are missing further testing. Bellow you can find more information on who we are, what cloud we are providing, what is OpenMinTeD and how is it related to Galaxy. Who are we ------------ I work for the Greek Research Network of Technology (GRNET S.A. [3]) as a software developer. The last few years I'm been involved in many cloud-related projects like Synnefo, ~okeanos and OpenMinTeD (more on them later). Synnefo, Pithos+, ~okeanos --------------------------- Synnefo [1] is an open source cloud software developed and supported by GRNET. It features an OpenStack REST API, many interesting features and it is around for more than 5 years already. Pithos+ is the object storage service of Synnefo, providing object storage with de-duplication, sharing and more. ~okeanos is the largest Synnefo deployment providing thousands of virtual machines and terabytes of storage to academic users in Greece and all over Europe. ~okeanos is also operated and supported by GRNET. OpenMinTeD ------------- OpenMinTeD is a collaborative project aiming in deploying text and data mining tools to academic communities. Galaxy ahs been selected as the main tool to support workflows. The plan is to run Galaxy on GRNET ~okeanos infrastructure, use Pithos+ object storage as a storage backend and make it collaborate with docker-oriented tools like Mesos/Marathon or Docker Swarm on ~okeanos provisioned VMs. [1] https://www.synnefo.org [2] https://www.synnefo.org/docs/synnefo/latest/pithos.html [3] https://www.openminted.eu [4] https://www.grnet.gr [5] https://www.synnefo.org/docs/kamaki/latest/ Stavros Sachtouris Software Engineer @ www.grnet.gr

1 0

Making symlinks in "command" section of tool XML
by Peter Briggs 19 Jan '17

19 Jan '17

Dear Developers I'm currently looking at develop a tool for a local pipeline, and want to make symlinks as part of the script in the "command" section of the tool XML - something like e.g. <command interpreter="bash"><![CDATA[ ## Create links to input FASTQs and Final_names.txt file #for $fq_pair in $fastq_pairs ln -s "${fq_pair.fastq_r1}" && ... #end for ... But when I try running the tool I get a failure with the message: Fatal error: Exit code 127 () bash: /PATH/TO/TOOL/DIR/ln: No such file or directory (NB I've snipped out the real path above, but essentially it's the directory where the tool XML is installed.) I know that it's permissible to use 'cp' and 'mv' in this context, so is 'ln' treated differently? Is there an alternative way to make symlinks in the command? Thanks for your help! Best wishes Peter -- Peter Briggs peter.briggs(a)manchester.ac.uk Bioinformatics Core Facility University of Manchester B.1083 Michael Smith Bldg Tel: (0161) 2751482

2 2