Comment on AWS accelerates instance reboots in anticipation of research findingsComments−ddorian438yCan't this be disabled somehow for internal servers only ? Say disabling it on db servers but keep it on api/web servers.−tomchukOP8yThey're doing this so you (or other shared tenants) can't read your VMs or the host's memory (and vice versa).−ddorian438yWhat about on dedicated servers ? Is there a config that can be done ?−rconti8yNo, and that's not the way the bug works.−Kudos8yThey're rebooting to patch the host, not your VM.−stephengillie8yThis is necessary because AWS does not support hot-migration of clients. To reboot the host, instead of pushing clients off to other hosts, they have to reboot all clients with the host.
Comments
Can't this be disabled somehow for internal servers only ? Say disabling it on db servers but keep it on api/web servers.
They're doing this so you (or other shared tenants) can't read your VMs or the host's memory (and vice versa).
What about on dedicated servers ? Is there a config that can be done ?
No, and that's not the way the bug works.
They're rebooting to patch the host, not your VM.
This is necessary because AWS does not support hot-migration of clients. To reboot the host, instead of pushing clients off to other hosts, they have to reboot all clients with the host.