my space suddenly went to a status “running” but it shows nothing except the spinning circle.
The problem is not my data in the root of my space, but my booked persistent space that seems to be tied to exclusively this one space. I can’t reach the subfolder to save or access my data from there.
what I tried so far:
1. the obvious:
restart
rebuild
did nothing. It says running all the time without changing anything
2. andvanced ideas
I renamed my app.py
I renamed my main*.py
I renamed the space (didnt work)
I tried to put it asleep (didnt work)
went into readme.md, tried an available gradio update, no dice.
So it’s “running” but IDK what to do now. I can’t duplicate it because there is no option to do so.
Can anyone help me out with a hint or where I can report my issue to restart the server?
Thank you for providing help and the mail address! I hope I did nothing wrong even tho I could not tell what it would be. It started with python, telling me about indentation errors when there were none. And after that, boom, it suddenly simply went dead.
I thought about switching to a pro subscription (was about to do so anyway) but I am not sure if this would motivate the server to re-initialize.
I contacted support and also the billing-team just to make sure nothing went wrong somewhere. I think I will go to sleep and try again tomorrow. thank you again for your support, this is very much appreciated!
@John6666, since you seem to know a lot about how this works, may I ask you one question:
if I delete the space and set it up again with the same name, will my persistent data in STORAGE_DIR = os.environ.get(“HF_STORAGE_DIR”, “/data”) survive?
(It’s not the viewable data folder that I can see in the root of the space).
I have never used Persistent Storage, so I cannot say for sure, but based on the behavior of other spaces, I think all data will be deleted when the space is deleted.
I think it will be necessary to ask HF for help in order to salvage the data. (If this is a bug, it will eventually be fixed since it is not normal behavior, and if it is some kind of flag, it should be possible to disable it with HF.)
meanwhlile I RTFM and asked Claude, 4o and DeepSeek to do some research. Turns out that persistent storage will last until the subscription is canceled.
So even if the space is deleted, the paid persistence storage will remain, until canceled.
However, since booked persistent storage is tied to exactly one space and not the whole account, it is necessary to delete the space and set up a new space with the exact same name (yes including CASE SENSITIVE!).
Since my only option is the HF support in any case, I think I will first follow your recommendation and
wait a few hours, hoping for a) the fix to disappear with server-maintenance or b) support taking care of my issue and my $5 payment VIP account.
try the nuclear delete-and-recreate option, if no support will happen. thumbs pressed.
Anyone else reading this for having similar issues: A recommendation (besides regular backups) is to use the dataset-storage. There is no CLI option for persistent storage due to security reasons.
It stays even after deletion! That’s good news. In the worst case, we can salvage the minimum code manually.
Well, let’s just hope that the space is restored.
Countermeasures against misuse are important, but it’s not good to cause users unnecessary effort and stress.
I 100% agree. I’m not even expecting too much support in my $5 case since I appreciate the HF low-budget options, but a good FAQ for beginners adressing exactly these issues would be very relaxing.
