#Object Stores (Globally availiable)
44 messages · Page 1 of 1 (latest)
In next 1-3 months. It is on our roadmap
This would be very nice to have, any updates @exotic sparrow?
Later today 👁️ (June 9th)
I don't have a time or anything, but I'll come mention you soon.
Just my luck this got rescheduled, we might need about a week
I lied - release today 😄
doesn't work:
same code worked with r2_id and r2_key instead
are you using your user id for auth?
nope, i used id and key from this view:
What if you use your runpod email as your ID?
nope you need use user id
Okay then update the docs primarily the code, for network storage s3 api.
Example code
ClientError: An error occurred (Unknown) when calling the ListObjects operation: signature mismatch for all known keys
why is this? just tried listobjects
using boto3
params = {
"Bucket": bucket_name,
"Prefix": prefix,
"MaxKeys": 1000,
}
if marker:
params["Marker"] = marker
resp = s3_client.list_objects(**params)
v2 also doesnt work, same error
okay this is weird Why wont it allow us to use access key ID from that
What happens if the disk is full?
well the request fails perhaps, it doesnt automaticaly expand
Storage capacity: Network volumes have a fixed storage capacity, unlike the virtually unlimited storage of standard S3 buckets. The CopyObject and UploadPart actions do not check for available free space beforehand and may fail if the volume runs out of space. This behavior is similar to applying a size quota in S3.
i hope runpod can allow the access_key_id generated instead of using the user id from secrets page
And im curious if
Failed multipart requests are billed
Like in AWS
You hav to remove the parts yourself
requests arent billed from the docs
oh i havent tested multipart yet
Does that eat up your storage
It should do multipart by default if you copy sth large
yeah i havent even uploaded any files yet using the s3 api hahah
Oof
Wut
Maybe too many ppl are using it
Finals week here
yeah probably, need to scale up
😢