3 ms·
Since day one, Amazon EC2 used a copy on write system with their LVM volumes to protect against this problem (without them having to do expensive zeroing operat
by comice 13y ago
Since day one, Amazon EC2 used a copy on write system with their LVM volumes to protect against this problem (without them having to do expensive zeroing operations).
This has been an identified and solved problem for YEARS. No excuse for a modern VPS/IaaS provider to be leaking customer data in this way, except incompetence.
- zimbatm 13y agoAmazon shifts the costs of zeroing on the user of the new VM. Because the blocks are zeroes on first access writes are twice as slow on the first pass. You can try that out easily by running `dd if=/dev/zero of=/mnt/test` twice.
- rpedela 13y agoYeah, but that isn't a big deal especially when you consider the alternative: leaked data.