A while back I ran into a configuration problem where the fs.conf file
accidentally listed the wrong handle range for a particular server (this
happened after the file system was created).
In this case, the trove handle management detects this problem because
the handles on disk don't seem to match the handle range in the conf
file. Unfortunately, the error message that tells you this is sent with
gossip_debug() rather than gossip_err(), so it is hard to tell what went
wrong without debugging enabled. This is the code snippet in
trove-handle-mgmt.c:
/* check every item in our range list */
if (!PINT_handle_in_extent_list(extent_list,
handles[i]))
{
gossip_debug(
GOSSIP_TROVE_DEBUG, "handle %llu is invalid "
"(out of bounds)\n", llu(handles[i]));
return -1;
}
Should this be changed to a gossip_err() instead?
On a related note, there are a lot of error messages in pvfs2-server.c
that just indicate general startup problems but are logged using
gossip_lerr(). Mostly they are just straightforward messages that tell
you that you set something up wrong. I would suggest that several of
these would be better off as gossip_err() rather than gossip_lerr(),
because the backtrace and line number information isn't especially
relevant at this level and it makes the startup errors look scary :)
-Phil
_______________________________________________
Pvfs2-developers mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-developers
