Botocore: pip install botocore: No matching distribution found for botocore==1.8.43

[codrutartene]

I have pip 9.0.1 (python 3.6) and I am trying to install gensim 3.3.0.

[andreineculau]

hitting the same issue now on travis with python 2.7

[racevedoo]

Same issue when installing awscli with python 2.7

[darabos]

Looks like 1.8.43 has been pushed incorrectly. It shows up on https://pypi.python.org/pypi/botocore/ but not on https://pypi.python.org/simple/botocore/.

[andreineculau]

friendly poke @JordonPhillips @kyleknap

[andreineculau]

related thread in aws-cli https://github.com/aws/aws-cli/issues/3149

[tiagocborg]

Oh, same issue here. Thank god you guys are fast on the responses.

[gekitsuu]

Any word on when we can see the fix?

[hekaldama]

@darabos I am new to python package publishing, I followed https://pypi.python.org/simple/botocore/ and see references to release 1.8.43 on there now. Does it seem like it was resolved or is it still missing something?

[jamesls]

Are people still seeing this issue? I'm not seeing issues when I try to install botocore. It might have been a transient issue with pypi as nothing about our release process has changed lately.

$ pip install --no-cache-dir botocore==1.8.43
Collecting botocore==1.8.43
  Downloading botocore-1.8.43-py2.py3-none-any.whl (4.1MB)
    100% |████████████████████████████████| 4.1MB 22.3MB/s
Collecting docutils>=0.10 (from botocore==1.8.43)
  Downloading docutils-0.14-py2-none-any.whl (543kB)
    100% |████████████████████████████████| 552kB 5.3MB/s
Collecting jmespath<1.0.0,>=0.7.1 (from botocore==1.8.43)
  Downloading jmespath-0.9.3-py2.py3-none-any.whl
Collecting python-dateutil<3.0.0,>=2.1 (from botocore==1.8.43)
  Downloading python_dateutil-2.6.1-py2.py3-none-any.whl (194kB)
    100% |████████████████████████████████| 194kB 15.0MB/s
Collecting six>=1.5 (from python-dateutil<3.0.0,>=2.1->botocore==1.8.43)
  Downloading six-1.11.0-py2.py3-none-any.whl
Installing collected packages: docutils, jmespath, six, python-dateutil, botocore
Successfully installed botocore-1.8.43 docutils-0.14 jmespath-0.9.3 python-dateutil-2.6.1 six-1.11.0

$ pip --version
pip 9.0.1

[hekaldama]

@jamesls let me try again. My issue is from an EMR cluster trying to run bootstrap scripts which update to the latest release of aws-cli. It seems to be intermittent in testing throughout the last couple of hours.

[darabos]

Yeah, it appeared to be fixed shortly after I saw the error. 🙂 Sorry for the panic. Perhaps I should have reported that it was back in working order for us.

So it seems like this was just a "natural" delay somewhere? 1.8.43 was missing for at least a few hours...

[jamesls]

Hmm, not sure. @dstufft any ideas?

[hekaldama]

@jamesls I have tested with > 4 clusters with _many_ nodes installing newest client and all are succeeding as they have in the past. This is resolved for me. Not sure on root cause and would be nice to know. Maybe we can get better mtimes on the uploads on pypi?

[codrutartene]

It is working well now!

[dstufft]

Yesterday the CDN provider on PyPI had a small disruption in the part of their service that handles PURGE requests. When the new version got uploaded to PyPI, and the PyPI backend software requested a PURGE of the currently cached entries, that purge didn't make it to every node in every PoP. This meant that some requests got served a stale response that didn't include the new release. This would have appeared to be an intermittent problem due to the nature of request routing within the CDN (you don't always get the same PoP or the same node within the PoP, and the differentiator between works or not is what PoP + node you were served by).

The disruption was quickly over, but it left any package that was uploaded in that time period in a potentially stale state. This would have fixed itself naturally over the course of a day or two as cached items fell out of the cache, however yesterday the PyPI admin team issued a manual purge to get everything back to 100%.