Giters

kubewharf / katalyst-core

Katalyst aims to provide a universal solution to help improve resource utilization and optimize the overall costs in the cloud. This is the core components in Katalyst system, including multiple agents and centralized components

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

memory advisor sendAdvices channel sometimes full

zhy76 opened this issue · comments

commented

What happened?

image

Maybe we need Increase capacity of sendChan

katalyst-core/pkg/agent/sysadvisor/plugin/qosaware/resource/memory/advisor.go

Line 89 in 6085bab

sendChan: make(chan types.InternalMemoryCalculationResult, 1),

What did you expect to happen?

no error

How can we reproduce it (as minimally and precisely as possible)?

get agent log

Software version

$ <software> version
# paste output here
commented

How often does this issue occur？

commented

It's kind of common on my side

commented

Actually, memory server will send response to qrm plugin when get advisorResp from memory advisor. If the channel is full, it indicates that the QRM plugin is overwhelmed with processing the results pushed by the memory server. If we simply increase the capacity of the channel, it will result in sending some outdated recommendation results to the QRM plugin, which is undesirable. Can we check if there are any abnormalities on the QRM side?

commented

I found qrm with these error:
image
image