49 Replies
403
Trying to start A1111 instance. Keeps saying "port not up yet"
?
Can't see the logs
try to open your own url in incognito mode
or private browsing
thats what other see
opens fine in incognito
Anyways, here.
ah
just my vpn blocked
yeah
its a problem with a1111, i think some other people has reported this
well damn.
it was working fine two days ago, same 1.10.0 template...
Its not a problem with A1111
Its a problem with installing multiple different A1111 templates on the same network storage and then they conflict with each other
Okay. I have just been using a newer template as that is what runpod automatically gives me. How can I fix this?
Which template are you using?
ashleykza/a1111:1.10.0.post7
Try that
Then connect to port 8000 and stop A1111 and start it again
run it in terminal inside jupyter ?
Correct, either Jupyter, Code server, Web Terminal, SSH etc
This filled up my storage disk even further. This is related to the issue I posted yesterday. I have no idea where 25-30GB of my storage has gone, and I keep needing to increase the storage disk size.
(which I obviously don't want to do since it costs money, for space that I have no idea why it is even filling up)
Seems to also not have fixed anything.
Can you DM me your port 7777 URL so I can take a look for you
This gives you access to the full pod? Preferably no.
Can't help then sorry
I have no interest in accessing your pod, was only trying to help, I fixed it for 2 other people
I understand, it just isn't all my own data which I am able to make the call on right now.
better just show the steps or explain it here
I don't know the steps, look above, I told him and it didn't work 🤦♂️
How about reading before replying
hey marcus its okay, don't need to be rude
I am not being rude, you are sending messages which are unnecessary and not adding any value
If you have nothing constructive to say, then don't say anything
maybe its because it needs to write cache on your container disk its temporary
I don't know what your affiliation with Runpod is, and from my perspective you are a random on the internet (no offense). I am also not dealing with only my own data, so I can't just give you access to the pod.
At this time, therefore, I cannot receive a solution which I cannot perform myself or through a verified runpod employee that is bound to their organization.
try deleting venv and template.json file and let it sync again, maybe you ran into issues when the syncing ran out of space
This may be it. Thank you for the idea.
Something like this
Then reset pod
At least this is clearing a ridiculous amount of space. Around 30Gb cleared already, and still going. My local install of A1111 is nowhere near that size.
well pip dependencies on the /venvs/ takes the most space i guess
It ended up clearing around 40Gb in total
Yep but probably also due to broken syncing when the disk space ran out
I don't think the venv should be 40GB in size, that doesn't seem right
which is a shame, since now I have a bunch of extra space to pay for every month that I have no need for.
(because I kept increasing network volume size constantly to keep up with the broken sync or whatever)
Fortunately the disk space isn't very expensive, like $7 a month for 100GB or something like that
yeah, silver linings and all that sure
GPU's are very expensive, so disk space doesn't really count in the grand scheme of things in my personal opinion
ChtGPT says the following based upon your errors:
It looks like you are encountering a SQLite database corruption issue while attempting to launch the stable-diffusion web UI. This error is related to the diskcache module, which is using SQLite for caching purposes.
It suggests you clear your cache directory: rm -rf /workspace/venvs/a1111/lib/python3.10/site-packages/diskcache/
The entire venv just got removed including that 🙂
oic
Claude 3.5 Sonnet beats ChatGPT by the way
yeah i mean sucks also to pay for an instance to troubleshoot lol
Oh yeah the time taken to troubleshoot must be really wasting your credits 😦
'tis what 'tis
Yep even though I hate that saying
I am just huffing some premium grade copium you kno
But I guess the costs come with the hobby. Sometimes I really understand the folks who never update a working environment.
Yeah you sometimes you pick stability over new and shiny
But other times you have no option when people like TheLastBen break their template 🙈
Llama 3.1: 😁🙏
Solved.
Removing venvs and letting it resync freed up a lot of space and fixed the malformed disk image issue.