[Help-gnunet] gnunet-check segfaults (broken gdbm files?)

[Markku Tavasti]

I have been running gnunet-insert something like 4 weeks
24h/7days/week. Did I mentition insertio is slow operation? (K6/350,
data and db on different disks, no much other load than gnunetd) :-(

Some insertions with -n, most of them without. Then comes storm, and
power starts flickering. Don't worry, I have UPS. But my scanner was
not connected to UPS, and and lost power while scanning some
paper. And box jammed. Gnunet-insert was still running, maybe only 10%
of whole job remaining.

Ok, reboot, no big problem. Start running gnunet-check. Starts fine,
founds errors since some files have been moved after indexing. But:

Bad content indexed for A5811360B40FB308B9EA0828BBD2B693FB0CCDF3
Will fix (deferred).
Segmentation fault  

real    1915m31.088s
user    240m36.800s 
sys     69m43.820s  

I'm running with default configs, gdbm database. I suppose it is
corrupted, what to do? 

a) A for fixing my db, do I run rm -rf and start over?
b) For fixing gnunet SW (Do I file a bug?)
c) Prevent this happening to me again. Is mysql more robust, and
maybe faster? Is it hard to set up gnunet with mysql? 


If some of the developers need my db, I can put it available (4.5G used)

-- 
M. Tavasti /  address@hidden  /   +358-40-5078254