On Tue, Jun 10, 2014 at 1:50 PM, John-Paul Robinson <j...@uab.edu> wrote: > Thanks for proposing a patch. > > We're looking at trying it out on our code base but are at least a few > revisions behind and may need to do a little back porting (or catching up).
I think it has been a while since this stuff has been significantly modified - I would guess the patch applies cleanly for last half dozen releases. > > Let me make sure I understand the intention of the patch, though. > > > >From the code changes, I trust that the intention is that all temporary > files should be created in the Galaxy dataset directory. That is the > temp directory path will be the explicitly derived from the output_path > that is an argument to upload.py. The attempt is for all temp files that will just be "moved" into output_path anyway be created and written out in the same directory as output_path. It is difficult to imagine scenarios where this is not on the same file system as output_path - this means shutil.move should be maximally performant and minimally error prone (agreed?). In general I agree that tools shouldn't really be modifying the TMPDIR - the system configuration should be used and the upload process may still result in some /tmp files - for instance uploading bam files causes some files for samtools stdout and stderr to be written to /tmp. This patch allow them to - because these files will never be "moved" to the output_path directory. > > This is fine. > > I'm somewhat partial to the environment influencing the TMP dir via an > unadorned mkstemp() but I can appreciate having a definitive > output_path specified on the call to upload.py. Responded to this above. > > One issue that will affect us slightly (but probably not others) is that > we currently have our dataset path on a glusterfs volume which doesn't > support ACLs. This means we won't have a way of overriding umask > settings that may be in place for the user. If a user with umask 022 > writes a file to their import dir, the permissions will at best be > read-only for the galaxy processes group owner. This may prevent with > the os.rename as well. With a default ACL we could cause the umask to > be ignored in favor of a permissive ACL. This is really a local deploy > issue but it's worth knowing about. So this expanded version of this pull request that will hopefully be included with the next Galaxy release in July should centralize some of the handling of this and have fewer paths through the code meaning it will be easier to tailor to your local deployment. At that time - if there is more we can do - adding different options, extension points, etc... to further ease your deployment let me know. > > A second question, it looks like the patch only updates convert_newlines > and convert_newlines_sep2tabs. The handle_compressed_file and sep2tabs > methods in sniff also have bare mkstemp() calls. Do they need > modification as well? Okay - so sep2tabs is used nowhere in Galaxy as far as I can tell. I would like to just delete it at some point. handle_compressed_file is never used during "upload.py" as far as I can tell - it is for formal data source tools - which should all behave more like normal jobs and normal Galaxy tools - so this is less likely to be a problem. There are some assertions in that last sentence I could be wrong about - and if that method proves problematic let me know and we can deal with it. Thanks again for the e-mail and for pushing Galaxy forward on this, -John > > ~jpr > > On 06/10/2014 10:49 AM, John Chilton wrote: >> Hello All, >> >> Thanks for the well laid out e-mails and great discussion. I think >> John-Paul's comment about the code growing up organically is probably >> exactly right. (A link below has some details from Nate about this). >> >> So late last night I opened a sprawling pull request that cleaned up >> a lot of stuff in upload.py and then realized it was a little bit >> incorrect and wouldn't actually help any of you until you were able to >> upgrade to the August 2014 release :) so I declined. I have reworked a >> relatively small patch to fix the immediate problem of the consistency >> tmpfile consistency as it relates to shutil.move. John-Paul Robinson, >> can you apply it directly and see if it fixes your problems? >> >> https://bitbucket.org/jmchilton/galaxy-central-fork-1/commits/dc706d78d9b21a1175199fd9201fe9781d48ffb5/raw >> >> If it does, the devteam will get this merged and I will continue >> with the upload.py improvements that were inspired by this discussion >> (see https://bitbucket.org/galaxy/galaxy-central/pull-request/408 for >> more details). >> >> -John >> >> On Mon, Jun 9, 2014 at 2:50 PM, John-Paul Robinson <j...@uab.edu> wrote: >>> We've considered the sudo solution, but it opens the window to other >>> bugs giving galaxy the power to change ownership of other files in our >>> shared user cluster environment. We could isolate the power to a script >>> but then we still need to monitor this code closely. We'd prefer not to >>> introduce that requirement. >>> >>> I didn't have the time to trace this down either. ;) I just got tired >>> of this issue and the inconsistent failures causing confusion in our >>> community. >>> >>> I hope your insight into the logic drift over time is accurate and can >>> be corrected. The upload code looks like it's gone through a whole lot >>> of organic growth. :/ >>> >>> Looking forward to additional comments from the dev team. >>> >>> ~jpr >>> >>> >>> On 06/09/2014 03:30 PM, Kandalaft, Iyad wrote: >>>> Hi JPR, >>>> >>>> I had the same questions while trying to figure out a fool-proof way to >>>> allow users to import files into galaxy on our Cluster. I couldn't >>>> exactly figure out, nor did I have the time to really review, why the >>>> galaxy code did these steps and why that shutil.move failed. I opted to >>>> simply insert code in upload.py to sudo chown/chmod the files as an >>>> "easier" "hack" to this problem. There are pros and cons to using the tmp >>>> var from the env, and it will depend on your intentions/infrastructure. I >>>> think the ideology was that the Galaxy folder was supposed to be shared >>>> across all nodes in a cluster, and they opted to use the TMP path within >>>> the galaxy folder. Overtime, the code probably partially diverged from >>>> that notion, which caused this dilemma. >>>> >>>> I believe that the best fix is to make the underlying code simply copy the >>>> files into the environment-provided temp, which is configurable in >>>> galaxy's universe_wsgi.ini, and assume ownership from the get-go. This >>>> code of copying and/or moving in discrete steps creates unnecessary >>>> complexity. >>>> >>>> Regards, >>>> Iyad >>>> >>>> -----Original Message----- >>>> From: galaxy-dev-boun...@lists.bx.psu.edu >>>> [mailto:galaxy-dev-boun...@lists.bx.psu.edu] On Behalf Of John-Paul >>>> Robinson >>>> Sent: Monday, June 09, 2014 3:08 PM >>>> To: galaxy-dev@lists.bx.psu.edu >>>> Subject: [galaxy-dev] inconsistent use of tempfile.mkstemp during upload >>>> causes problems >>>> >>>> There appears to be some inconsistent use of tempfile.mkstemp() within >>>> upload.py that causes problems when users import data files to galaxy from >>>> a cluster directory via the upload process and import/temp/dataset >>>> directories are on different file systems. >>>> >>>> The issue manifests when Galaxy's job directory, dataset directory and >>>> import directory are on different file systems (common for cluster >>>> environments) in conjunction with a configuration where users can copy >>>> their data files directly to the import directory from which Galaxy >>>> selects data sets to upload (as opposed to using an FTP gateway). >>>> >>>> While allowing users to copy files to an import directory rather than >>>> using the FTP gateway may not be that common, we use this configuration >>>> locally to help build a more seamless interface with our local collection >>>> of HPC resources. Users can be logged into their cluster account and move >>>> data into galaxy with a file copy command rather than having to use FTP. >>>> >>>> This configuration has worked well in our environment as long as the >>>> correct ownership configuration existed on the import directory and as >>>> long as the import directory, job temporary directory, and galaxy data set >>>> directory were all on the same file system. >>>> >>>> We now have our galaxy dataset directory on a different file system and >>>> are seeing inconsistent behavior during the upload.py runs depending on if >>>> the data is ordinary text, BAM files, or gzipped data. >>>> >>>> A subset of uploads will fail because of the way temporary files are >>>> created by Galaxy to facilitate the import and any associated conversion >>>> processes of different file types. >>>> >>>> During the import, >>>> >>>> 1) Galaxy will copy the original file to a temporary target file >>>> (converting as needed during the copy). >>>> 2) Once this first conversion step is complete, Galaxy then attempts to >>>> move the temporary file back to the original location, ie. the import >>>> directory. >>>> 3) If this move is succeeds, Galaxy completes the upload processing and >>>> the data becomes a registered data set in the user's dataset collection. >>>> >>>> Galaxy prefers the Python shutil.move method to move tempfile . This >>>> results in a simple os.rename if the files remain on the same file system. >>>> However, if os.rename raises OSError because a move was attempted across >>>> a file system boundary, shutil.move resorts to a copy2, which copies the >>>> data to the original import file and then tries to copy the file >>>> attributes (permissions and utimes) to the original import file from the >>>> source file (which will be the temporary file Galaxy created in step 1 to >>>> begin the conversion process). >>>> >>>> The os.rename and shutil.copy2 behave (and fail) differently depending on >>>> the file ownership of the original import file. The os.rename will >>>> succeed even if the Galaxy upload.py job process only maps to the >>>> group-owner of the original import file (which can be ensured with group >>>> sticky bit on the import dir or ACLs). The shutil.copy2 command, however, >>>> will fail if the Galaxy upload.py job process UID is not the user-owner of >>>> the original import file. >>>> >>>> We could ensure the os.rename succeeds by keeping the job temporary >>>> directory and the import directory on the same file system. However, it >>>> seems the temporary directories used by upload.py are inconsistent across >>>> data types which prevents this simple fix from working for all data types. >>>> >>>> When text files are imported, upload calls the sniff.* methods to perform >>>> conversion. These methods use a bare call to tempfile.mkstemp() which >>>> ensures the file is created in the directory specified by the env var >>>> $TMPDIR. For example in sniff.convert_newlines: >>>> >>>> fd, temp_name = tempfile.mkstemp() >>>> >>>> https://bitbucket.org/galaxy/galaxy-central/src/5884328e91724e9bdf4b43f012eb63fa0c803ef6/lib/galaxy/datatypes/sniff.py?at=default#cl-105 >>>> >>>> However, for compressed files, the upload.py script directly creates temp >>>> files but here it specifies the target directory as the same as the data >>>> set directory: >>>> >>>> fd, uncompressed = tempfile.mkstemp( >>>> prefix='data_id_%s_upload_gunzip_' % dataset.dataset_id, >>>> dir=os.path.dirname( output_path ), text=False ) >>>> >>>> https://bitbucket.org/galaxy/galaxy-central/src/5884328e91724e9bdf4b43f012eb63fa0c803ef6/tools/data_source/upload.py?at=default#cl-130 >>>> >>>> It's not clear if there is any significance to using the data set >>>> directory as the tempdir for compressed files versus the job temporary >>>> directory for other data files. >>>> >>>> It seems like all temporary files created by upload.py should be >>>> consistently created in the same temporary location, and preferably in the >>>> job temp directory. >>>> >>>> Is there a reason that these file types use different temporary file >>>> locations? >>>> >>>> If they used the same tempfile location, we could use one consistent >>>> system configuration and ensure all our data files can be imported even >>>> when the import+tempdir are not on the same file system as the Galaxy >>>> dataset dir. It seems reasonable that all tempfile.mkstemp() calls should >>>> be unadorned and inherit the temp directory location from their >>>> environment. >>>> >>>> A more comprehensive solution that would correct the inconsistency in >>>> failures between os.rename and shutil.copy2 and also remove any constraint >>>> for Galaxy to have it's import, temp, and data set directories on the same >>>> file system, would be to simply delete the original import file before >>>> attempting the shutil.move. This would ensure the file that the upload.py >>>> job attempts to create in step 2 is new and created with full Galaxy >>>> process ownership. >>>> >>>> Finally, it seems odd that Galaxy attempts to reuse the users original >>>> import file in the first place. It seems that once galaxy begins >>>> processing the content of the to-be-imported file, it should not ever >>>> write back to that file. What's the motivation here? >>>> >>>> I'll be interested to learn more about the motivations of these different >>>> tempfile conventions and if this can be fixed in the upstream. >>>> >>>> Thanks, >>>> >>>> ~jpr >>>> ___________________________________________________________ >>>> Please keep all replies on the list by using "reply all" >>>> in your mail client. To manage your subscriptions to this and other >>>> Galaxy lists, please use the interface at: >>>> http://lists.bx.psu.edu/ >>>> >>>> To search Galaxy mailing lists use the unified search at: >>>> http://galaxyproject.org/search/mailinglists/ >>> ___________________________________________________________ >>> Please keep all replies on the list by using "reply all" >>> in your mail client. To manage your subscriptions to this >>> and other Galaxy lists, please use the interface at: >>> http://lists.bx.psu.edu/ >>> >>> To search Galaxy mailing lists use the unified search at: >>> http://galaxyproject.org/search/mailinglists/ > > ___________________________________________________________ > Please keep all replies on the list by using "reply all" > in your mail client. To manage your subscriptions to this > and other Galaxy lists, please use the interface at: > http://lists.bx.psu.edu/ > > To search Galaxy mailing lists use the unified search at: > http://galaxyproject.org/search/mailinglists/ ___________________________________________________________ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
