Connection Reset Browser


We are having issues uploading through odrive. I am able to create folders, but all uploads fail, less than 50% of the way.

Chrome Dev tools shows “net::ERR_CONNECTION_RESET” as the reason for the upload failing.

The drive i’m using is backed by google cloud. However, I am able to upload just fine when using the Google Cloud Console directly.


Any update? This is causing 100% of uploads to fail.

It is not a wifi or local network connection issue.

Hi @keith,
There doesn’t appear to be a general issue as I am able to upload to my Google Cloud Storage accounts.
Are you testing with both the web client and the desktop client?
Are you seeing this with all files sizes?
Are you able to download?

Hi @keith,
Quick update. I am seeing there may be an issue on one of our nodes that could be causing issues with uploads via the web browser, although the desktop client should be fine.

We have taken that node out of rotation, for the time being. Let me know if you are still seeing the issue.

Files sizes are mainly ~2-3MB.
I’ve only tested with the web client.

The first download failed, connection refused.

I then did a DNS cache flush, and now its working… o.0 … Maybe there was a bad server or the ip changed recently?

Thanks for taking a look Tony. I’ll add a computer restart to the things to try again in the future.

Ah. My second test must have lined up with you taking out that node. Uploads look fine now.

Thanks Tony!

On a side note… are there any extra support features or priority on the premium accounts? That’s not something I can’t find, but would pay for…

Hi @keith,
If it was the webclient it may have been this specific node you were hitting. Routing is based on latency, so it may have been the one you were using. When we took it offline you may have still been hitting it, which is why your downloads started to fail too. We’re taking a look at it now.

Thanks for the report!

For support, it is something we are looking at (expanding the support department and “off hours” response time). My apologies for the delay on this one. The way this node failed was subtle, so it wasn’t picked up by our automated monitoring.