Dear All,
we are testing out a simple, single server xrootd/rootd setup
and it looks like using rootd leaves two defunct processes per
rootd session:
xrootd 18972 18961 0 14:35 ? 00:00:00 [rootd <defunct>]
xrootd 18974 18961 0 14:35 ? 00:00:00 [rootd <defunct>]
xrootd 19327 18961 0 14:37 ? 00:00:00 [rootd <defunct>]
xrootd 19328 18961 0 14:37 ? 00:00:00 [rootd <defunct>]
one at the beginning of the rootd connection, the second showing up
after the client completes. The log file does not report any problem:
080716 14:35:47 18961 XrdProtocol: ?:14@rhenium terminated link transfered
080716 14:35:47 18961 XrdProtocol: ?:14@rhenium terminated link transfered
080716 14:37:12 18961 XrdProtocol: ?:14@rhenium terminated link transfered
080716 14:37:12 18961 XrdProtocol: ?:14@rhenium terminated link transfered
The rootd part of the config looks like:
xrd.protocol rootd /usr/local/xrootd/lib/libXrdRootd.so \
/usr/local/xrootd/bin/rootd -i -r /cdf/local/disk \
-r /cdf/local/scratch -r /export/data -r /cdf/disk
I see the problem of the defunct processes only when the server runs
on an SL 3 machine (SL 4 leaves no defunct processes, identical setup
and clients).
Is this a know problem, is there a work around for this? Thanks,
cheers, Stephan
|