rhettg / Tron

Next generation batch process scheduling and management
Other
11 stars 0 forks source link

Machine failure could potentially leave empty state file #74

Closed rhettg closed 12 years ago

rhettg commented 12 years ago

We had a hard failure on our tron machine.

Though there is pretty good resiliancy around writing out new state files, we still managed to come back up after reboot with 0 bytes.

The theory is that even though the file is closed, that doesn't actually do an fsync so we're not guaranteed it's written out before overwriting.