Mountpoint Updates Jan 2026 and Feedback Wanted!

[adpeace]

Hi Mountpoint Community!

We recently turned on Discussions to create a dedicated space for community feedback and sharing use cases that help us prioritize development. For those new to the project, Mountpoint for Amazon S3 is a high-performance file client that lets you mount S3 buckets as local file systems on Linux: check out our README and product page for more details.

We wanted to start the conversation by sharing some updates on recent changes and improvements to Mountpoint and what we’re currently working on.

Over the past year we’ve released many improvements to make Mountpoint more usable at scale:

• Mountpoint is now available directly in the Amazon Linux 2023 repository, so you can install and patch to the latest update more easily. For information on how to install and update on Amazon Linux 2023, please see the Amazon Linux 2023 section of the installation guide.
• We added support for mounting buckets via fstab so you can automatically mount your S3 buckets at boot and persist the mount across boots, as well as set up mountpoint as part of an EC2 userdata script. For information on the options available and example fstab entries, please see the Configuration guide.
• We added metrics and support for exporting these via the OpenTelemetry Protocol for better observability in your existing monitoring stack. The Metrics documentation describes how to publish metrics from Mountpoint and details of the metrics that are being published.
• We improved sequential read performance by approximately 40% with two or more network interfaces on our multi-NIC benchmarks with large objects (40GB downloads), measured on the trn1.32xlarge in us-east-1. See our multi-NIC benchmark configuration in the benchmarks directory.
• Lastly, we improved memory efficiency. You can now see the memory usage during our regularly-run benchmarks here: the changes have reduced variability in these graphs and peak memory usage by over 50% in all but one benchmark.

For Kubernetes users we also launched version 2 of our Mountpoint for Amazon S3 CSI driver that accelerates performance for repeatedly accessed data, adds support for Security-Enhanced Linux (SELinux) mount options, and simplifies logging and permissions management.

We’d love to hear how you’ve been using these new improvements – please reply if you have feedback on these changes, or file a GitHub issue if you’ve encountered any problems using them.

What’s next?

We have a number of projects upcoming and will be reviewing/updating our public roadmap over the coming weeks. One major area we are continuing to work on is memory management as we still hear this feedback from customers. We're working on adding a memory limit option to let you set explicit memory budgets. This gives you direct control over the performance/memory tradeoff. If you're running Mountpoint in resource-constrained environments or have feedback on memory usage in v1.20, we're keen to hear your feedback.

Feedback and contributing

Let us know how you're using Mountpoint, e.g. what workloads are you using with Mountpoint – ML training, data analytics, or something else? Are there any issues you've experienced or changes you’d like to see in Mountpoint that would make your workload or its operation smoother?

Please share feedback by replying to this post, opening a new discussion, or opening an issue or pull request if you have a specific problem or request. We look forward to hearing from you.

Thanks for being part of the Mountpoint community!
– The Mountpoint team.