GalaxyFAQ: Difference between revisions

From Cheaha
Jump to navigation Jump to search
(→‎Error: channel 3: open failed: connect failed: Connection refused: Removed the emphemeral port comment. These ports are dynamically selected so shouldn't interfere.)
(→‎When I hit Execute from Galaxy, does that mean the job is running on the cluster?: Add information and steps involved in the run states of a job launched from galaxy web interface)
 
(4 intermediate revisions by the same user not shown)
Line 4: Line 4:
== Personal Galaxy ==
== Personal Galaxy ==


=== Error: channel 3: open failed: connect failed: Connection refused ===
=== What does ''channel 3: open failed: connect failed: Connection refused'' mean? ===


This error shows up when attempting an ssh port forwarding from the galaxy server to the localhost. This error occurs due to port conflict on the server side. The port number mentioned during the execution of ''personalize.sh'' might already be taken up ie., a service might already be running at this port. Check the list of ports opened by doing a ''netstat -lpn |grep <port-number>''. If the result is not empty, run the ''personalize.sh'' script again with a higher port number and then ssh port forward with this new port number. On Unix/Linux systems, normal users can open ports starting from 1025 to 65535.
This error shows up when attempting an ssh port forwarding from the galaxy server to the localhost. This error occurs due to port conflict on the server side. The port number mentioned during the execution of ''personalize.sh'' might already be taken up ie., a service might already be running at this port. Check the list of ports opened by doing a ''netstat -lpn |grep <port-number>''. If the result is not empty, run the ''personalize.sh'' script again with a higher port number and then ssh port forward with this new port number. On Unix/Linux systems, normal users can open ports starting from 1025 to 65535.


=== I've set up a Personal Galaxy, now what? ===
=== I've set up a Personal Galaxy, now what? ===
You could run jobs either from the command-line or from the galaxy web interface.
# Command-line - Run jobs in either [https://dev.uabgrid.uab.edu/wiki/GalaxyUsingToolWrappersFromCommandLine#Interactivemode interactive mode] or [https://dev.uabgrid.uab.edu/wiki/GalaxyUsingToolWrappersFromCommandLine#Non-interactivemode batch mode] on the SGE cluster.
# Web interface - Follow either the [[Galaxy_DNA-Seq_Tutorial|DNA-Seq]] or [[UAB_Galaxy_RNA_Seq_Step_by_Step_Tutorial|RNA-seq]] tutorials to run some NGS QC and manipulation jobs on the cluster from the galaxy web interface. The datasets related to these two tutorials can be downloaded from the [https://galaxy.uabgrid.uab.edu/ UAB Galaxy server]  from its [https://galaxy.uabgrid.uab.edu/library shared data libraries.] NOTE: Access to UAB's galaxy server requires blazer-id authentication.


=== When I hit ''Execute'' from Galaxy, does that mean the job is running on the cluster? ===
=== When I hit ''Execute'' from Galaxy, does that mean the job is running on the cluster? ===
Yes. Almost all ''Execute'' operations for the tools listed in the left-hand side of the galaxy web interface result in a running a ''qsub'' job on the cluster. Once you hit the ''Execute'' button, the job goes through a series of events such as:
# For the first few minutes after execute, you will see the watch icon next to the dataset in the History pane on the right hand side. This means that the job has been submitted to the queue and is in queue wait/hold state. Monitor this activity on the cluster by doing a ''qstat''.
# After a while, you will see a spinning icon next to the dataset. This means the job is running on the cluster.
# Once the spinning icon disappears, this means that the job is complete. You may perform any of the following actions to view your results:
## View the results in the data browser (middle pane) by clicking on the ''Eye'' icon in the History pane, to the right of your dataset.
## Click on the finished dataset itself. This displays a menu below the dataset in the History pane. The displayed options are to save, view details, run the job again, edit, annotate tags associated with the job.
### If you click on the ''info'' icon, you can view details on the job run like, the standard output, standard error and the actual file path on the cluster where your resulting datasets are located.


=== Where are the log files (stdout/stderr) for the job executed from Personal Galaxy? ===
=== Where are the log files (stdout/stderr) for the job executed from Personal Galaxy? ===
Galaxy provides two ways in which you can access the job's stdout and stderr content.
:1. database - Galaxy stores the standard out and standard error in the database. The table which contains job's stdout and stderr is ''job'' and to view the same, use the following sql query
<pre>
sqlite> select tool_id,stdout,stderr from job where id=23;
tool_id        stdout                                stderr   
-------------  ------------------------------------  ----------
fastq_trimmer  2913558 fastq reads were processed.
</pre>
:2. web interface - Galaxy web interface provides an html link to the job's stdout and  stderr. This can be accessed by clicking on the tool panel's (the panel on the righthand side) info button for each tool -> Tool Standard Output / Tool Standard Error links.
:During the run of the job on SGE runner, the .o (stout) and .e (stderr) files are created in the directory specified by ''$file_path'' in universe_wsgi,ini. These .o and .e files are removed automatically, once the job is complete. To preserve these two files on the file system, run galaxy in debug mode ie., set ''debug = True'' in universe_wsgi.ini.

Latest revision as of 20:06, 11 June 2012

FAQ for Galaxy

Personal Galaxy

What does channel 3: open failed: connect failed: Connection refused mean?

This error shows up when attempting an ssh port forwarding from the galaxy server to the localhost. This error occurs due to port conflict on the server side. The port number mentioned during the execution of personalize.sh might already be taken up ie., a service might already be running at this port. Check the list of ports opened by doing a netstat -lpn |grep <port-number>. If the result is not empty, run the personalize.sh script again with a higher port number and then ssh port forward with this new port number. On Unix/Linux systems, normal users can open ports starting from 1025 to 65535.

I've set up a Personal Galaxy, now what?

You could run jobs either from the command-line or from the galaxy web interface.

  1. Command-line - Run jobs in either interactive mode or batch mode on the SGE cluster.
  2. Web interface - Follow either the DNA-Seq or RNA-seq tutorials to run some NGS QC and manipulation jobs on the cluster from the galaxy web interface. The datasets related to these two tutorials can be downloaded from the UAB Galaxy server from its shared data libraries. NOTE: Access to UAB's galaxy server requires blazer-id authentication.

When I hit Execute from Galaxy, does that mean the job is running on the cluster?

Yes. Almost all Execute operations for the tools listed in the left-hand side of the galaxy web interface result in a running a qsub job on the cluster. Once you hit the Execute button, the job goes through a series of events such as:

  1. For the first few minutes after execute, you will see the watch icon next to the dataset in the History pane on the right hand side. This means that the job has been submitted to the queue and is in queue wait/hold state. Monitor this activity on the cluster by doing a qstat.
  2. After a while, you will see a spinning icon next to the dataset. This means the job is running on the cluster.
  3. Once the spinning icon disappears, this means that the job is complete. You may perform any of the following actions to view your results:
    1. View the results in the data browser (middle pane) by clicking on the Eye icon in the History pane, to the right of your dataset.
    2. Click on the finished dataset itself. This displays a menu below the dataset in the History pane. The displayed options are to save, view details, run the job again, edit, annotate tags associated with the job.
      1. If you click on the info icon, you can view details on the job run like, the standard output, standard error and the actual file path on the cluster where your resulting datasets are located.

Where are the log files (stdout/stderr) for the job executed from Personal Galaxy?

Galaxy provides two ways in which you can access the job's stdout and stderr content.

1. database - Galaxy stores the standard out and standard error in the database. The table which contains job's stdout and stderr is job and to view the same, use the following sql query
sqlite> select tool_id,stdout,stderr from job where id=23;
tool_id        stdout                                stderr    
-------------  ------------------------------------  ----------
fastq_trimmer  2913558 fastq reads were processed.
2. web interface - Galaxy web interface provides an html link to the job's stdout and stderr. This can be accessed by clicking on the tool panel's (the panel on the righthand side) info button for each tool -> Tool Standard Output / Tool Standard Error links.
During the run of the job on SGE runner, the .o (stout) and .e (stderr) files are created in the directory specified by $file_path in universe_wsgi,ini. These .o and .e files are removed automatically, once the job is complete. To preserve these two files on the file system, run galaxy in debug mode ie., set debug = True in universe_wsgi.ini.