I’ve been searching for material regarding LVS’ failover status detection
mechanism, and I was wondering if anyone could refer me to material that can
answer the fallowing questions:
When a Linux Director makes a decision whether a node is alive or not, what
does this decision is based on?
Is it base on a collection of statuses such as hardware statuses (e.g. .disk
space, cpu usage etc.), processes statuses (e.g. up, running etc.) ?
Is there any place in which I can configure such kind of status parameters?
If I wish to measure a parameter that can’t be configured - how can I
integrate the code that checks it into LVS?
-Thanks
-Miri
|