Re: Server Error Trying To Dl Data From Galaxy On Cloud

Question: Re: Server Error Trying To Dl Data From Galaxy On Cloud

7.7 years ago by

Hello, I'm trying to download the library files I processed on my galaxy cloud instance, but I'm getting an error. At the top (on the right panel) it says "Server Error" and then lists the URL where the data should be and then lists: Module paste.exceptions.errormiddleware:143 in __call__ << try: __traceback_supplement__ = Supplement, self, environ app_iter = self.application(environ, start_response) return self.make_catching_iter(app_iter, environ) except:>> app_iter = self.application(environ, start_response) Module paste.debug.prints:98 in __call__ << try: status, headers, body = wsgilib.intercept_output( environ, self.app) if status is None: # Some error occurred>> environ, self.app) Module paste.wsgilib:544 in intercept_output << try: for item in app_iter: output.write(item) finally: if hasattr(app_iter, 'close'):>> output.write(item) MemoryError: out of memory Is there some easy fix to this? I'd really like to get that data off of the cloud instance and be able to terminate it. thanks, karl

galaxy • 1.0k views

ADD COMMENT • link •

modified 7.7 years ago by Enis Afgan • 680 • written 7.7 years ago by karlerhard@berkeley.edu • 90

7.7 years ago by

Enis Afgan • 680

Enis Afgan • 680 wrote:

Hi Karl, As you see from the error message, you seem to be getting this error because the machine is running out of memory. This can in part be caused by a configuration option that might be set in Galaxy's universe_wsgi.ini file (see below). Did you have any jobs running while trying to download the file? Waiting until those finish might free up some memory. A thing to try is to connect to the instance, edit Galaxy's universe_wsgi.ini file to se debug = False, restart Galaxy and try again. Are you familiar with that at all? The basic steps are as follows: [local]$ ssh -i <path to="" your="" aws="" private="" key="" file=""> ubuntu@<instance public="" ip=""> [ec2]$ sudo su galaxy [ec2]$ cd /mnt/galaxyTools/galaxy-central [ec2]$ vi universe_wsgi.ini -- edit file (around line 226) to set: debug = False [ec2]$ sh run.sh --stop-daemon [ec2]$ sh run.sh --daemon Yet another option is to connect the instance in the same way, look through the data library on the file system and manually copy the file out of the instance. You can use the following command to copy the file to your local instance: [local]$ scp -i <path to="" your="" aws="" private="" key="" file=""> ubuntu@<instance public="" ip="">:/mnt/galaxyData/files/000/dataset_<id>.dat . Let us know if none of this works, Enis

ADD COMMENT • link written 7.7 years ago by Enis Afgan • 680

Hello Enis, Thanks for the quick response and suggestions. I actually did have a job running while I tried to download a file the first time, that's the first time it gave the error message. But the jobs have long since finished and it's still giving the error message. I've been able to edit the universe_wsgi.ini file to "debug = False", but now I'm getting an "Internal server error" when I try to reload the galaxy instance. Should I be signed out at the command-line to reload galaxy from a browser? Forgive my simplicity, I'm really not at all command- line savvy. Also, another extremely basic problem I have is I just don't know where the data library that holds my files is located. Any help would be greatly appreciated. best, karl

ADD REPLY • link written 7.7 years ago by karlerhard@berkeley.edu • 90

Hi Karl, Hmm, not having Galaxy accessible is definitely not a step in the right direction. Being signed into command line is not an issue; something else must have gone wrong. To start, please take a look at the (bottom of) galaxy log file (and email the relevant part if you don't see how to fix it immediately); the file is saved as /mnt/galaxyTools/galaxy-central/paster.log As far as the location of the files in data libraries, they should be stored in the same location as history datasets, namely /mnt/galaxyData/files/000/dataset_<id>.dat Because all of the datasets are named simply based on the database ID, it won't necessarily be obvious which file to get without doing some (python) coding or doing some guess work. If you know how large your files is, you can easily narrow your choices down by listing the contents of the given directory and sorting it by size (using command ls -lS), then pulling out the file(s) that you want. If several files are of approx. the same size, open them up and see which one you want. Good luck and let us know if you have any more trouble, Enis

ADD REPLY • link written 7.7 years ago by Enis Afgan • 680

7.7 years ago by

Enis Afgan • 680

Enis Afgan • 680 wrote:

You're almost there, the command should be executed from your local machine (home directory is fine) and it should look as follows: scp -i <path to="" keyfile=""> ubuntu@<publicdns>:/mnt/galaxyData/files/000/dataset_11.dat . (note the 'ubuntu@' before the <public dns=""> and a trailing dot (.) - the dot means your current directory on your current machine, i.e., your home dir if that's where you are executing the command from) I apologize for the trouble in getting the this data out and hope it does not keep you from using Galaxy Cloud in the future (we're looking into why the browser-based data copy didn't work and should have a fix shortly for the main app). Enis

ADD COMMENT • link written 7.7 years ago by Enis Afgan • 680

Excellent, it was the trailing dot that I was missing! Thanks so much for the help, I will most certainly be using Galaxy again, it's been very useful so far. karl

ADD REPLY • link written 7.7 years ago by karlerhard@berkeley.edu • 90

Please log in to add an answer.

Similar posts • Search »