I was running inside a screen session so the logging vanished off the top 🙄
I should add some actual logging to the application...
Lars Holm Nielsen
@lnielsen
Ok….I’ve found the failed request in the logs
Graeme Winter
@graeme-winter
Cool - there is a very good chance that the issue was on the interconnect or at my end... but adding some more resilience would be a good thing I think
Lars Holm Nielsen
@lnielsen
I can see it send a 502 bad gateway from our end
and that it managed to go to ~5GB uploaded first
does it fit that it was uploading for some 15-20 minutes before failing?
Graeme Winter
@graeme-winter
Yeh that fits
anything I can do at my end to be more resilient e.g. restart the upload at the failed point?
Lars Holm Nielsen
@lnielsen
unfortunately we don’t do resumable uploads :S
we’re looking into it and trying to debug it…..right now,it’s just to try again….which is not very nice for a 20 minute upload :(
_
Graeme Winter
@graeme-winter
I'll leave with you for a little while then - for the moment it seems to be fine with files <= ~ 1GB or so which should cover most of our use case. Thanks for looking into it :)
Graeme Winter
@graeme-winter
@lnielsen we are going now - looks like people are appreciating this too https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind2003&L=CCP4BB&D=0&P=190211 - thanks again for your help. Quick question, probably for Monday - currently reading API is in development according to your pages - would it be possible for you to give me some hints how I could make a script which will perform a keyword search and then grab the title / doi / file list for each? This will mean I can liaise with the other data archive people (the protein data bank) to allow linking back to the data sources...
Lars Holm Nielsen
@lnielsen
great news!!! Thanks for keeping us in the loop. If I get a moment during the weekend, I’ll try to send you info on the api
OK, followed up on this finally - all good, got what I was looking for 🙂 thank you
Graeme Winter
@graeme-winter
Good morning 🙂 now trying to update automatically some of these records to (i) fix a typo in a keyword and (ii) add a couple of extra words to the descriptions - weird thing I am finding is that it works sometimes but now appears to be failing with an invalid action
{'message': 'Invalid action', 'status': 403} Traceback (most recent call last): File "/Users/graeme/git/zenodo-uploader/zenodo_updater.py", line 203, in <module> updater() File "/Users/graeme/git/zenodo-uploader/zenodo_updater.py", line 198, in updater zenodo_updater.update() File "/Users/graeme/git/zenodo-uploader/zenodo_updater.py", line 132, in update self._update() File "/Users/graeme/git/zenodo-uploader/zenodo_updater.py", line 106, in _update raise RuntimeError("in update: HTTP status %d" % r.status_code) RuntimeError: in update: HTTP status 403
At around 07:13 UTC
Lars Holm Nielsen
@lnielsen
Checking
Graeme Winter
@graeme-winter
👍 thank you
Lars Holm Nielsen
@lnielsen
Do you have the deposit id that it happened to?
Graeme Winter
@graeme-winter
Located deposition: id = 3730327
Lars Holm Nielsen
@lnielsen
Usually this happens if e.g. a record was not put in editing mode, or is already in editing mode…..