Question: What is the maximum supported cluster size for QFS?

[zhanglistar]

Hi Team,

I'm considering QFS for a large-scale storage infrastructure and would like to understand its maximum supported cluster size.

Background:

• Target cluster: 1000+ storage nodes
• Expected raw capacity: 10+ PB
• File count: Millions to billions

Questions:

1. Node limits

   • Maximum chunkservers per cluster?
   • Maximum metaservers? Is multi-master or metaserver federation supported?
   • Any limits on clients concurrently accessing the cluster?

2. Storage limits

   • Maximum total raw capacity (theoretical and practical)?
   • Maximum size per file?
   • Maximum size per chunkserver?

3. Metadata limits

   • Maximum number of files/directories?
   • Maximum namespace size? Any inode limits?
   • How much memory does the metaserver typically need per million files?

4. Performance considerations

   • What becomes the bottleneck at very large scale (metaserver CPU/memory, network bandwidth, disk I/O, or Zookeeper)?
   • Are there any known scalability issues or bugs when exceeding certain thresholds?

5. Production experience

   • Has anyone deployed QFS at 1000+ node scale in production?
   • What's the largest cluster size known to be running QFS successfully?

I've reviewed the documentation but couldn't find specific scalability numbers. Any insights, benchmarks, or references to real-world deployments would be extremely helpful.

Thank you for your time!

P.S. If these limits are documented somewhere, please point me to the link. Much appreciated!