xor wrote: > This looks like a physical corruption of your Freetalk database. > Are you running it on a network drive or any other sort of unreliable > disk? Notice that old harddisks can have bad sector errors. > Was the system hard-reset (power failure)? > > It seems that you are on Linux, so please obtain smartmontools and do a > full test of your harddisk: > smartctl -t long /dev/<harddisk> > > It will show you how long the test takes. > > After than, you can get the result via: > smartctl -a /dev/<harddisk> > > The self-test results are usually at the end of that.
Ummm. I will try that, but the box has more uptime than the launch of freebox (7 days now, it is a headless server), and the node has been shutdown always properly in these seven days. In any case, is there some file I should delete? Anything named freetalk in the node folder? I have ./freetalk_data.db4o ./freetalk-testing-17.db4o ./Freetalk/Freetalk.db4o Well, I went ahead anyway and deleted the three of them, and FT loaded. So it was indeed database corruption. Could a node dying of OOM cause this? It's the first that comes to my mind (and SMART status is OK).
