. How can you choose where a crash dump will be generated?
. What are common avenues to explore if the crash dump shows that the node ran out of memory?
. What should you look for if the process count is suspiciously low?
. If you find the node died with a process having a lot of memory, what could you do to find out which one it was?
Using the analysis of a crash dump in Section 6.1:
. What are specific outliers that could point to an issue?
. Does it look like repeated errors are the issue? If not, what could it be?