- This issue started presenting itself after the security upgrades for meltdown.
- This is suspected to be a performance issue due to such security update.
- We know that due to meltdown, postgresql is 40% less performance.
- MAAS runs with all 5 processes (4 regionds, 1 rackd).
- There's 5 VM's running on the same system (details unknown), that seem to be booting at the same time.
- The system has 10 cores.
So, due to basic analysis, this issue could be a direct result of having the VM's contending for resources + MAAS using high resources as well as other processes, and the metldown patches.
So far, through our investigation, we have not identified any specific issue in MAAS to fix in order to improve/avoid this issue.
As IRC discussion over last week.
- This issue started presenting itself after the security upgrades for meltdown.
- This is suspected to be a performance issue due to such security update.
- We know that due to meltdown, postgresql is 40% less performance.
- MAAS runs with all 5 processes (4 regionds, 1 rackd).
- There's 5 VM's running on the same system (details unknown), that seem to be booting at the same time.
- The system has 10 cores.
So, due to basic analysis, this issue could be a direct result of having the VM's contending for resources + MAAS using high resources as well as other processes, and the metldown patches.
So far, through our investigation, we have not identified any specific issue in MAAS to fix in order to improve/avoid this issue.