ah. you are right - I had not updated that older repository - I think I was waiting for confirmation that the testtoolshed version worked - will do shortly - thanks for pointing that out! Try removing that one completely and then installing the test toolshed version please? That should work until I get the main one updated in a few hours. On Mon, Dec 2, 2013 at 9:28 AM, Karen Miga <khmiga@soe.ucsc.edu> wrote:
Yes, I do all of my installs from the toolshed through the admin menu, however I did not use the "test toolshed..." I used the main. Ah ha - maybe that was the issue?
Karen
On Sun, Dec 1, 2013 at 2:23 PM, Ross <ross.lazarus@gmail.com> wrote:
Just to confirm, you used automated installation via the admin menu -> search and browse toolsheds -> test toolshed -> etc ? If you installed manually by downloading and unpacking the archive under your tools/ directory, then yes, there is no automation and you need to install the binary on your path. I believe the documentation explains this - if it's not clear please suggest some revisions ?
On Mon, Dec 2, 2013 at 9:18 AM, Karen Miga <khmiga@soe.ucsc.edu> wrote:
Hi Ross,
That is listed as "tool_dependency_dir = /mnt/galaxy/tools" in my universe_wsgi.ini file, and that did not contain a clustalw directory either before or after the toolshed install.
Karen
On Sun, Dec 1, 2013 at 2:10 PM, Ross <ross.lazarus@gmail.com> wrote:
yes. In universe_wsgi_ini you have a definition for tool_dependency_dir - that's why I wrote: [tool_dependency_directory]/clustalw/2.1/fubar/clustalw/76de6971a0de/ as the path - the funny numbers are the repository revision number - that's where the toolshed installs the binary and the env.sh file needed for the tool to find it.
On Mon, Dec 2, 2013 at 9:08 AM, Karen Miga <khmiga@soe.ucsc.edu> wrote:
Hi Ross,
I did not detect any type of auto-install (?) After installing from toolshed I did not find the binary under the /mnt/galaxy/shed_tools/ toolshed.g2.bx.psu.edu/repos/ (or /mnt/galaxy/tools or /mnt/galaxy/tools/bin for that matter). Am I looking in the wrong place?
Karen
On Sun, Dec 1, 2013 at 2:00 PM, Ross <ross.lazarus@gmail.com> wrote:
Hi Karen, Glad to hear it worked for you. It the tool auto-installs correctly, it downloads and unpacks the binary under [tool_dependency_directory]/clustalw/2.1/fubar/clustalw/76de6971a0de/ so the binary you installed manually should not ever be needed by the toolshed version - try moving it off your path - the tool should continue to work correctly? Apologies - I should have suggested removing any manual installations and their corresponding tool_conf.xml entries before using the automated installation - the tool shed installation doesn't prevent you from having 2 versions of the same tool in your tool menu - until those manually installed ones (which may depend on clustalw2 being on the path) are removed?
On Mon, Dec 2, 2013 at 8:36 AM, Karen Miga <khmiga@soe.ucsc.edu>wrote:
> Hi Ross, > > Thanks for your email. I have clustalw2 working now - I installed > the clustal repository (as suggested) from toolshed and then installed the > clustalw2 binaries in our mnt/galaxy/tools/bin. > > Thanks again! > Karen > > > On Wed, Nov 27, 2013 at 12:46 AM, Ross <ross.lazarus@gmail.com>wrote: > >> Hi, Karen - how did you install clustalw? >> If not already done, can you please try installing the clustal >> repository owned by fubar from the test toolshed through your Galaxy admin >> interface - search/browse toolsheds? Please let me know how you go and if >> there's a problem I'll try to fix it for you. I think there's a clustal >> omega tool too but it's not one of mine. As far as I know my clustal >> repository should install cleanly on a CloudMan instance. >> >> Sorry but we'll have to wait to from someone who knows about HyPhy. >> >> Hello - I am running to errors when trying to run Clustalw and >> HyPhy on our cloudman instance. >> >> (1) Clustalw: The program completes without error and produces an >> empty file and a log file that states "/bin/sh: 1: clustalw2: not found" >> >> Searching around a bit I am not seeing clustalw2 previously >> installed. I am happy to do it if need be, however, I wanted to check in >> and see if I am overlooking the directory or an easy fix. >> >> If I do need to install clustalw, I would prefer to use >> clustalw-omega, but was not sure if I needed to script a brand new wrapper >> or if I could use the previous one that came with galaxy image. >> >> (2) HyPhy: The cloud the link was initially broken. I changed it >> to the correct path: >> >> sudo ln -sfn /mnt/galaxy/tools/hyphy/default >> /mnt/galaxy/galaxy-app/tool-data/HYPHY >> >> I am using a fasta alignment file of 6 protein sequences (566 AA >> each) as a test case (*also fails with nucleotide alignments). >> >> It runs briefly and then issues an error statement "Single >> Alignment Analyses Segmentation fault (core dumped)". >> >> Thank you in advance, >> >> Karen >> >> >>
-- Karen H. Miga, PhD
Postdoctoral Scholar University of California, Santa Cruz (831)459-5232
-- Ross Lazarus MBBS MPH; Head, Medical Bioinformatics, BakerIDI; Tel: +61 385321444 http://scholar.google.com/citations?hl=en&user=UCUuEM4AAAAJ
-- Karen H. Miga, PhD
Postdoctoral Scholar University of California, Santa Cruz (831)459-5232
-- Ross Lazarus MBBS MPH; Head, Medical Bioinformatics, BakerIDI; Tel: +61 385321444 http://scholar.google.com/citations?hl=en&user=UCUuEM4AAAAJ
-- Karen H. Miga, PhD
Postdoctoral Scholar University of California, Santa Cruz (831)459-5232
-- Ross Lazarus MBBS MPH; Head, Medical Bioinformatics, BakerIDI; Tel: +61 385321444 http://scholar.google.com/citations?hl=en&user=UCUuEM4AAAAJ