Could you please let me know why do we need ReadWriteMany volume on k8s for deploying openreplay? what it store inside that volumes? We actually use on-prem cluster which does not have RWX volume, so kinda stuck here.
We’ve some workloads that need it. One which writes the ingested data to a file (sink), second one (storage) reads the processed data, then does some optimizations (sorting and compression) before uploading to the object storage. Both need to be scaled differently, as per your need. That’s why we need the RWX volume.
If you don’t have RWX, you can still add nodeSelectors, and taints to group sink/storage/chalice in the same node, and use hostPath as volume. But the system won’t be resilient. For that, you really need a RWX volume.
Having said that, that data is temporary cached in that volume, so unless you’ve a huge traffic, it shouldn’t be an issue, as storage is fast enough to process the data as it’s ingested. But there’s a risk you end up with a few corrupted sessions in case any issue occurs.