Good catch. Patch is committed. We fixed this once before but it must
have unfixed itself somewhere along the line :)
As a side note, in testing this I happened to try a configuration file
that had a completely bogus line in it (ie, a bad keyword, not just a
bad parameter). Despite the patch, this case still allows the server to
start because of a different logic decision at dotconf.c:762. In that
case it just warns about an unsupported config file option and keeps going.
Anyone have any thoughts on if that is the right approach in that case?
It seems reasonable if the config file option is a (for example)
optional tuning parameter that isn't present in different PVFS versions,
but not so much if the line represents something critical to correct
operation. Unfortunately I don't think there is a way to differentiate.
-Phil
Nicholas Mills wrote:
Hey all,
I found a bug in the error checking for PINT_parse_config(). For some
reason this function returns 0 on succes and 1 on error. The current
code in pvfs2-server.c throws an error if the return value is less
than zero, which obviously never happens. As a result the server
doesn't shut down after a configuration error.
I made a few simple changes to fix the problem. They are included in
the attached patch file.
--Nick
------------------------------------------------------------------------
_______________________________________________
Pvfs2-developers mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-developers
_______________________________________________
Pvfs2-developers mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-developers
