Getting Bad HTTP Status Code error from some Instances
Opened this issue · 5 comments
I'm getting this error from some of the instances but not all of them:
func = lambda : node.run_tasks(opts, args, conf)
File "/usr/local/lib/python2.7/dist-packages/brenda/node.py", line 333, in run_tasks
proj_dir = get_project(conf, blender_project)
File "/usr/local/lib/python2.7/dist-packages/brenda/node.py", line 424, in get_project
get_s3_project(conf, url, proj_dir)
File "/usr/local/lib/python2.7/dist-packages/brenda/node.py", line 382, in get_s3_project
file_len, etag = aws.s3_get(conf, s3url, fn, etag=etag)
File "/usr/local/lib/python2.7/dist-packages/brenda/aws.py", line 92, in s3_get
return paracurl.download(dest, url, **paracurl_kw)
paracurl.Exception: (-5, 'PC_ERR_HTTP : bad HTTP status code')
Probably my settings, but I'm having trouble figuring it out.
Any ideas?
I was able to re-run a similar job without this error. Seems as though I was trying to use subfolders in my buckets. The tutorial does specifically prescribe against this, but since the S3 console let me add subfolders, I thought it would be okay. Turns out you need to use a separate bucket for source and output areas, and no subfolders. Lesson learned!
It would be nice in the future to somehow support sub folders, especially if you want to further automate the render job launching process, but it's not biggie right now.
I am getting the same problem now (-5, 'PC_ERR_HTTP : bad HTTP status code') on all machines and I haven't messed with a sub-folders. Can anyone show me direction what could I check to find where have I mistaken. Thanks.
I solved the problem by a installing on a client:
apt-get install python-pip gcc python-dev libcurl4-openssl-dev git
so some of those was making trouble probably.
I had this same error. Turns out I had my project file name misspelled in .brenda.conf...
I saw this error as well, and it was because I accidentally pushed the .blend file to s3 instead of the .tar.gz