Details

Reviewers

Group Reviewers

Commits

rDLDBASEd1f35f5231a9: package.loader: add a test to reproduce EOFError error

Summary

This exception has been seen on staging but we were unfortunately not
able to reproduce it. It could be related to underlay storage
issues. Here is the traceback:

Traceback (most recent call last):
File "/usr/lib/python3/dist-packages/swh/loader/package/loader.py", line 291, in load
  self._load_revision(p_info, origin)
File "/usr/lib/python3/dist-packages/swh/loader/package/loader.py", line 368, in _load_revision
  uncompressed_path = self.uncompress(dl_artifacts, dest=tmpdir)
File "/usr/lib/python3/dist-packages/swh/loader/package/loader.py", line 211, in uncompress
  uncompress(a_path, dest=uncompressed_path)
File "/usr/lib/python3/dist-packages/swh/core/tarball.py", line 72, in uncompress
  shutil.unpack_archive(tarpath, extract_dir=dest)
File "/usr/lib/python3.7/shutil.py", line 999, in unpack_archive
  func(filename, extract_dir, **kwargs)
File "/usr/lib/python3.7/shutil.py", line 934, in _unpack_tarfile
  tarobj.extractall(extract_dir)
File "/usr/lib/python3.7/tarfile.py", line 2002, in extractall
  numeric_owner=numeric_owner)
File "/usr/lib/python3.7/tarfile.py", line 2044, in extract
  numeric_owner=numeric_owner)
File "/usr/lib/python3.7/tarfile.py", line 2114, in _extract_member
  self.makefile(tarinfo, targetpath)
File "/usr/lib/python3.7/tarfile.py", line 2163, in makefile
  copyfileobj(source, target, tarinfo.size, ReadError, bufsize)
File "/usr/lib/python3.7/tarfile.py", line 247, in copyfileobj
  buf = src.read(bufsize)
File "/usr/lib/python3.7/bz2.py", line 178, in read
  return self._buffer.read(size)
File "/usr/lib/python3.7/_compression.py", line 68, in readinto
  data = self.read(len(byte_view))
File "/usr/lib/python3.7/_compression.py", line 99, in read
  raise EOFError("Compressed file ended before the "
EOFError: Compressed file ended before the end-of-stream marker was reached

This happened when decompressing the file
https://downloads.sourceforge.net//asio/asio-1.12.1.tar.bz2 but the
loader executed locally can decompress it without any issue.

Diff Detail

Repository

rDLDBASE Generic VCS/Package Loader

Branch

catch-EOFError

Lint

No Linters Available

Unit

No Unit Test Coverage

Build Status

Buildable 11256
Build 17010: tox-on-jenkins	Jenkins
Build 17009: arc lint + arc unit

Event Timeline

lewo created this revision.Mar 20 2020, 11:29 AM

Herald added a reviewer: Reviewers. · View Herald TranscriptMar 20 2020, 11:29 AM

lewo added a subscriber: ardumont.Mar 20 2020, 11:29 AM

Build is green
See https://jenkins.softwareheritage.org/job/DLDBASE/job/tox/370/ for more details.

Harbormaster completed remote builds in B11242: Diff 10173.Mar 20 2020, 11:30 AM

Did you try truncating a .tar file to reproduce the issue?

Or rather, a .tar.bz2

ardumont added inline comments.Mar 20 2020, 12:45 PM

swh/loader/package/loader.py
381	you can add a test on functional loader which patches the `uncompress` method to raise this.

In D2859#68753, @vlorentz wrote:

Did you try truncating a .tar file to reproduce the issue?

No, and that's a good idea!
I just tried but unfortunately, I didn't reproduce;(

I actually succeeded with a bigger truncated file actually: 70KB

I added a test for the EOFError.

Fix a typo

Build is green
See https://jenkins.softwareheritage.org/job/DLDBASE/job/tox/375/ for more details.

Harbormaster completed remote builds in B11255: Diff 10186.Mar 20 2020, 4:15 PM

Build is green
See https://jenkins.softwareheritage.org/job/DLDBASE/job/tox/376/ for more details.

Harbormaster completed remote builds in B11256: Diff 10187.Mar 20 2020, 4:17 PM

D2860 and especially are going your way D2862 ;)

ardumont mentioned this in D2860: package.loader: Reference a snapshot on partial visit.Mar 23 2020, 9:16 AM

What about this diff? I think it's still valuable since it catch more precisely errors and add a test case;)
WDYT?

Yes, i think we can land it after D2860 and D2862 ;)

D2862 will removes the try: except in the _load_revision and captures all in the main loop.
So in effect, that will make that diff add a scenario and not touch the code.