- Two pools, both running XS 6.5 + latest patches as of 11/9/15
- Storage is dedicated iSCSI 20Gbit LACP bond
- Pool A consists of HP Bl460c Gen8 Blades
- Pool B consists of HP BL460c Gen9 Blades
- Example VM in this case is a small (500MB) Debian Wheezy used for templates
- Migrate VM on pool A to pool B
- Start VM on pool B and write to disk
- Reboot VM
- Filesystem corruption!
If you don't notice this before writing much data to the VM the filesystem is destroyed and cannot be recovered (easily).
The migration process causes the the disk to change in a manor that results in EXT4's free block count to change. (See screenshot)
If this is not immediately repaired and data is written to the filesystem corruption will occur.