howdy all - just starting to get familiar with Pin...
# getting-started
j
howdy all - just starting to get familiar with Pinot. I’ve got a REALTIME table feeding off of kafka for testing purposes and that’s it. I’m pumping a lot of data at for stress testing. One thing I’m not understanding is why my controllers have their disk space filled? I know controllers are involved somehow in dealing with offline segments, but I’ve not configured this yet. From the controllers logs:
Copy code
Caught exception while uploading segment: aws_flowlogs__0__331__20220714T1655Z from instance: Server_pinot-server-5.pinot-server-headless.pinot-eng.svc.cluster.local_8098
java.nio.file.FileSystemException: /var/pinot/controller/data/pinot-controller-1.pinot-controller-headless.pinot-eng.svc.cluster.local_9000/fileUploadTemp/aws_flowlogs__0__331__20220714T1655Z.ef130a54-cefe-4619-91b4-e1370c086d4f -> /var/pinot/controller/data/aws_flowlogs/aws_flowlogs__0__331__20220714T1655Z.tmp.73e9073d-2d5d-4e92-899d-8a246177f759: No space left on device
thanks for any explanation to this newbie!
👀 1
x
hey Jeff!!
j
How Xiaobing 🙂
I meant hi!
x
Pinot comes with a
deepstore
by default I think it’s using controller’s local disk
in usual prod setup, the deepstore is pointing to s3 or other kinds of remote storage system
lemme pull out the wiki for your reference
j
I see. If it’s on by default, then that would obviously explain it…
j
I’d not gotten to the Deep Store arena yet - I’ll give a look! Thanks so much, and really nice to chat!
😉 1