It happened again. I delete an Rpub document, but rpubs claims slug is already taken

I reported this in the Fall of 2020, but no help.

Students complained that they were getting an 404 error for an html document generated by knitr from rstudio.

Somehow it was deleted from my profile https://rpubs.com/briholt

In particular, there should be a documetn called "getting started" with a slug "Getting_started".

But it's not there.

So how can I delete it? And if I go to that space https://rpubs.com/briholt/Getting_started it shows that it's broken.

I'm in a bit of a catch 22 and so advice would be appreciated

Sorry! I have no idea how that happened, for now I have corrected the problem by hand--you should be able to publish now and reuse that slug.

Thank you. I will test it soon. I have another document with the same problem.

Is there a way for me to fix it so you don't have to waste your time doing it manually?

No--and for the life of me I can't figure out how it got into this state. The document was in the database in what seems like an impossible state: assigned to your account, slug set, but not yet claimed by a user (and thus not eligible to be displayed). If the Getting_started one works, tell me the other slug and I'll reset it.

It did work! Thank you!

The other file, which is still broken: http://rpubs.com/briholt/Abnormal_schedule

Is it possible that publishing from different computers could cause this?

My process for the abnormal_schedule was that I would make changes to the markdown file, from my home linux machine, push commits to github. Then at work on a windows machine, would pull the changes, make some edits, commit, push, but then publish from that windows machine. That was about "when" students noticed that the link no longer worked.

For the the getting_started file you just fixed, I think I published from a different linux workspace on my home computer. When I did that, students began to squak that it wasn't working.

I don't know how any of this works, so I'm really just speculating. It could all have been coincidental.

At the time, I began to wonder if the .rpoj file stored the slug, or something.

Thank you for your time and help.

Oh, that certainly seems like a helpful hint. Are either of those git repos public by any chance?

They are public, I believe

OK, I fixed Abnormal_schedule. Still haven't found any more clues about why this happened, but I'll keep looking.

Did either of these pubs possibly fail to publish at some point? i.e. at first it successfully published, but during some subsequent to update it, you got an error of some kind, and it maybe even updated successfully after that?

Ah, yeah, I figured it out. If there's an error during republish of an existing pub with a slug, you get into this broken state. I'll see about fixing it.

Thanks!

2 Likes

Thank you so much for your help. How were you able to track this down? Is there a log somewhere? Probably one of your servers, I imagine. I'm wondering if there is a way to search for these errors. I wonder how often this happens.

By the way, in December I received an email from a user who had the same problem.

I mentioned this post here and suggested that they make a new help ticket and to link it here.

Oh, last comment, I don't recall receiving any notices about an error.

Dear All,

On December 17, 2020, I had a similar problem with two documents. For the first 6-7 months, everything was OK: I updated the documents without any problem. And then, suddenly, I got an error during updating. After that, I could not update the files and finally they disappeared sending a 404 error message, when I tried to open. My attempts to create new documents with the same slug failed.

I was really upset, since I already shared the links with many people.

I would be very grateful to Mr. Cheng, if he could fix my problem too. In general, it would be nice to prevent such problem from occurring in the future, so that no more people face it again.

Best

I didn't find anything helpful in logs, but just looking at the "impossible" state in the database was enough of a clue for me to just pore through the code and reason it out.

It looks like in the ~8.5 years since RPubs launched, this has happened 218 times, though only 117 of those times did the document have a slug (and thus present a problem that cannot be worked around).

Interesting. Thank you for sharing it. Clearly a huge problem that rstudio should put all of their resources to fix. Of course I'm kidding.

Thank you so much for your help.

@Konstantin_Holodilin Please email me at joe@rstudio.com and I'd be happy to help you out. If you include the links that are broken I can reset them.

The good news is that I've identified the problem in the code and have a potential fix, hope to get it out in a release in the coming weeks.