We recently got notifications for a couple of our instances, one retirement and one reboot. I stopped and started the instances today, and the instances did not come back up. Looking at the "system log" output, I saw some kernel output from the guest's kernel (ending at "Speculative Store Bypass: Vulnerable"), then some Xen output "Pagetable walk" and "domain_crash_sync called from entry.S: fault at XXXXX entry.o#compat_create_bounce_frame+0xa1/0xed". I didn't change anything on the instances, I figure this has to be a bug in the xen update that presumably necessitated the maintenance. AWS wants to force me to pay to report this directly, so posting on a forum in hopes somebody there will see it. It looks like this is not happening on t1.micro (yet)
AWS 官方已更新 2 年前