We have a pretty urgent/critical issue affecting our day-to-day operations, and I was wondering if any of you faced a similar issue before.
We have a few automations that are copying some fields from one table to another. In our specific case - it’s attachments.
Since today, we have started seeing very odd issues - when the file gets copied from one table to another (Via an automation), another Automation that has no name (see screenshot attached) delets the file attached.
A few facts
The only change that I did was that I started testing Stacker, however I disabled the “Share with Link” feature on my base and it didn’t change anything.
Welcome to the community, @Albert! :grinning_face_with_big_eyes:
Regarding the automation that’s copying the attachments from one table to another, is it also deleting the old attachment from the former table in the step immediately following the one that copies it? If so, that probably the cause.
When adding attachments, Airtable needs time to fully transfer the file to its own storage and do other processing before it’s technically fully attached. It’s easy to assume that the transfer is instantaneous, but it’s not. If the original file is removed before the transfer is complete, it aborts the attachment and removes the partial file, which is what you’re seeing.
My recommendation: delete the automation step that removes the attachment from its original table, and set up some other automation process to batch-remove those attachments at a later time.
Thank you for your reply.
It’s not deleting the files from the former table, and there is no step like that. The files are kept in the other table forever.
Also, it seems like today it started working again. It would be nice to know about these things from the Airtable team though, meaning that there is an issue they are working on, so that the users wouldn’t spend the whole day debugging ¯_(ツ)_/¯
Sometimes the issues are just as much a surprise to them as they are to you. Not sure how their “alert the masses” process works, but it’s possible that for some issues they may just dive into fixing it rather than tie up others with sending an alert that might be moot by the time it gets out.