Migrated Instance, Cannot Upload to Amazon S3 Anymore
jimmybrancaccio opened this issue · comments
Hi all,
I recently migrated my personal pasteboard to a new server. I literally just tar.gz'd the directory and then SCP'd it over to a new system. I installed it requirements, unpackaged the tar file and ran ./run_local script. Upon attempting to upload a file, it won't actually upload to Amazon S3. It will show in the pre-upload state but once I click on the Upload button, it will attempt to upload, go to the next page but show Image Not Found.
I was running this on some old code, so I attempted to just run a fresh install and did a new git clone. I reconfigured it all it looks like Bitly is no longer used so the only auth/ file I had worked with was the amazon.js one. I just re-added my keys and info (bit confused on the CDN URL though), and still same issue. It's not uploading it to S3. I tried using local storage and that worked fine. I straced and ran a tcpdump on the processes and it's definitely connecting to Amazon but I cannot seem to determine if they're giving any error or anything. I checked and confirmed my S3 bucket permissions are good. I even went back to the old installation and tested an image upload and it worked fine, still went to S3, so it's got to be something about this server or install or something. @JoelBesada Any ideas or tips I could use to troubleshoot? I thought there was some logging on this but it was another app I had migrated so I don't seem to have any logs for this specific app to refer too. I'll probably answer this question myself in a minute, but is there a debug mode to put the app in?
Any help is much appreciated.
P.S. I thought I would have been annoyed that the Bitly/shorturls were gone, but I'm not, I like how it displays the embed URL, I had hacked around my older code install to show this since that's the URL I really needed since I always upload a picture to post on my blog or something, but now it does by default 👍
It just magically started working. It's possible it was either the region of the bucket or the CDN_URL which I had set as http instead of https.
I wanted to quick update this because it happened again. I determined the issue was being caused by the clock being off. Just ran ntpupdate -b -u pool.ntp.org to sync the servers clock and all was happy and well again.