All my serverless instances are "initializing" forever
I am trying with my own template (from the tutorial) and trying to spin up a canned template (the Llama) but I just get an endless "Initiailizing" (see image). Can someone help? This is really holding me back.
25 Replies
Try to set the max worker to 2. They will initialize up to 5 workesr for you, but you don't pay for them, only the ones that are active, and they will honor your max worker.
See if that ends up fixing it. Sometimes Ive seen this fix this issue.
Hmmm. Okay. I tired 1, 2, and 3. Just spins on initializing. How long should it take? I’d imagine not more than a minute or two right?
Yeah, if you continue to have an issue, take a picture of the template that you used + post your pod ID here
The staff is best to help you
or maybe you can kill the end point by deleting it and then trying again
But I would say that im surprised, that you are having issues rip. Definitely though just provide more info on the template + the endpoint ID and the staffs can also check the logs post-afterwards to see why. But maybe killing the endpoint and trying something else should be ok.
Yeah, I've removed and reinstalled numerous times. Both canned templates (e.g. the Lllama one) and my own from the tutorial.
Here's the screenshots
FYI it's not a pod, it's serverless
here's the console from the instance
You are missing a tag for your docker image in the template
And its also a bad idea to use the
latest tag, be sure to tag your docker image with a version so you can easily deploy new versions when you need to.Okay. Good best practices. But I don't think that's what the issue is here. this is just the tutorial docker so the 'latest' shouldn't impact too much here. I did add a tag when I created this. Let me make that edit.
Where would I put a 'tag' in in the template? I don't see that option:
Sorry... here's serverless
Its still not specifying a tag for the docker image
container image should specify a tag like
:1.0.0Where would I do that? I don't see that as an option
Its a text box, not an option, set image to
tmattoneill......:1.0.1 for example instead of tmattoneill......Again, this is helpful best practice. But I have the same issue with the templated ones, too. Here's the Llama one, which has been initiaitalizing for 10+ minutes with no activity. Is it user error? (possible for sure). But I can't get my head around why this isn't working.
Do you have any requests in the queue?
I run the request at the bottom; it spins for 2-3 minutes, then says Network Error (trying the Llama Template as a test)
May you provide a screenshot of this?
Hey guys. So there was an issue of some sort on your end. I've changed nothing in my config but it's working now. started working again yesterday evening GMT.
I now have "running" and can use the platform.
I had the ‘endlessly initializing’ thing last night, working through the tutorial. Turns out I’d forgotten to push the docker image. 👀
Same here, I've created a new serverless pod and it's still initializing after hours of creating it
That could be a problem with your docker image, are your workers throttled or not?
Right, It was infact a problem with the docker, it turns out the docker image I'm using is not on docker hub. I wish there's a way for me to check the logs in serverless with this kind of issues
You can click on a worker and check its logs for these issues.
Ohh! actually, I was using your docker image, to run exllama2 in runpod.