Wei-Chiu Chuang created HDDS-14797:
--------------------------------------
Summary: [Docs] Service limits
Key: HDDS-14797
URL: https://issues.apache.org/jira/browse/HDDS-14797
Project: Apache Ozone
Issue Type: Wish
Components: documentation
Reporter: Wei-Chiu Chuang
I'd like to have a page that describes the Ozone Service Limits. Similar to
this AWS doc (Quotas and service limits for AWS Organizations):
[https://docs.aws.amazon.com/organizations/latest/userguide/orgs_reference_limits.htm]
Using Gemini CLI, I managed to come up with a number of missing service limits.
Missing Documentation in Apache Ozone
1. General Scalability and Cluster Limits
The documentation does not specify the tested or theoretical limits for:
* Maximum number of datanodes per cluster.
* Maximum number of users.
* Concurrent Access: Maximum number of simultaneous file reads or writes per
user or for the cluster.
2. Bucket Limits (Compared to S3)
Inspired by S3's documented limits, the following are not specified for Ozone
buckets:
* Maximum number of buckets per volume or system.
* Bucket name length (min/max characters).
* Bucket policy size limit.
* Bucket ACLs: Maximum number of grants per ACL.
* CORS Configuration: Maximum number of rules.
* Lifecycle Configuration: Maximum number of rules.
3. Object / Key Limits (Compared to S3)
Similarly, for objects (or "keys" in Ozone terminology), the following limits
are not documented:
* Maximum number of files/objects per bucket.
* Maximum object/file size.
* Object key/file name length (max characters/bytes).
* Maximum object size in a single PUT operation.
* Object Tags: Maximum number of tags per object.
* Object ACLs: Maximum number of grants per ACL.
* Object Versioning: Maximum number of versions per object.
4. Multipart Upload Limits (Compared to S3)
The limits for multipart uploads are not specified:
* Number of Parts: The maximum number of parts per upload.
* Part Size: The minimum and maximum size for a part.
5. Performance Benchmarks
* The documentation provides guidance on performance tuning but lacks
concrete benchmark numbers (e.g., IOPS, throughput, latency) to set performance
expectations.
6. Hardware Recommendations
* While there are specific recommendations for network and disk hardware,
there are no specific CPU or RAM recommendations for the different service
roles.
7. Backup and Disaster Recovery
* There is no dedicated guide detailing the strategy and steps for backing
up and restoring an entire Ozone cluster. Existing disaster recovery guides
explicitly state that backup procedures are out of scope.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
