Hi, We have 30 ignite clients connecting to 1 ignite server and all of them are creating multiple cache and upon completion of their tasks the cache are destroyed. This destruction triggers a Partition Map Exchange and initiates binary metadata cleanup for each cache. During this period we start encountering BufferUnderflowException, and storage becomes temporarily unavailable. Note: we do not encounter this issue with 10 client. This has been observed in Ignite version 2.17 How to overcome this? -Rohan