#error starting: Error response from daemon: Container aa58de3216b8515a3ee78aa46d9102331aaaf6c210a36c

58 messages · Page 1 of 1 (latest)

spark pike
#

Hey all, all my requests stopped working on p45dj1lfott9ob.

Example: 2319a9c9-dccd-4930-b34f-0acce67bdc3c-u1
Example: 40c01e77-32c9-451e-ac7a-dc134b513a97-u1

Can someone helped me look into why this is happending? Thanks

simple relic
#

request id wont help

#

Whats up, any logs?

#

or errors

#

(in text please)

spark pike
#

there's no logs. all requests are put into delay state.

simple relic
#

delay?

#

no running worrkers?

spark pike
simple relic
#

did you you use any from quick deploy?

simple relic
spark pike
#

it just delay infinitely

simple relic
spark pike
#

there's no logs

simple relic
spark pike
#

and the worker is in idle state

simple relic
#

I don't know whats wrong, did your code ever worked before?

spark pike
#

yes

simple relic
#

it just happened today or now?

spark pike
#

it just happened suddenly with that error message

simple relic
#

after worker running that, it stopped, and you sent request again it didn't run again?

spark pike
#

yes

simple relic
spark pike
#

it's custom rendering pipeline

simple relic
#

How do you call your serverless.start()

#

in your code

spark pike
#

if name == 'main':
runpod.serverless.start({
"handler": predict,
"return_aggregate_stream": True,
})

#

async def predict(job: Dict):
// code

#

even I killed the worker, the container loading seems to get stuck

simple relic
#

No its not stuck

spark pike
#

it has been like that for minutes now. It usually shows the docker pulling progress bar

simple relic
#

its still loading

#

Thats downloading

#

its only loading

spark pike
#

ok

simple relic
#

thats normal...

spark pike
#

sorry I killed the worker after I could not figure out what's going on

simple relic
spark pike
#

RP web interface -> terminate

simple relic
#

maybe* it was running and you killed it

#

okay..

spark pike
#

i don't know it was running. It's not taking any requests

simple relic
#

oh you only have 1 workers?

#

try 3 max workers

spark pike
#

yeah it's a dev server

simple relic
#

tell me after, when it loads

#

use 3 anyways

spark pike
#

I can't becaue the max 5 limit

#

oh nvm, I can

#

Trying it now

whole heath
#

Even if you have a max 5 limit, add credits and increase your worker count, you should never set max workers to 1 UNDER ANY CIRCUMSTANCES.

#

RunPod set the default to 3, don't change it!

spark pike
#

got it.. thank you all. I also suggested maybe a warning message to prevent users not setting it below 3...

#

otherwise, I think new users might miss it

whole heath