It sounds like there is some corruption going on after some use. Are
you saying that transfers work initially, but then you get this error
from uberftp, but then you can restart uberftp and transfers will again
work initially?
If this is the case, would it be possible to run the client under
valgrind? http://valgrind.org/ It is a tool that will detect memory
corruption and other problems that might be causing this. It is fairly
simple to use. After building, I would just run it like this (one line):
valgrind --log-file=/tmp/uberdebug --undef-value-errors=no uberftp
<uperftp-opts>
And then use uberftp normally until the error occurs, and send the log file.
Mike
Osamu Tatebe wrote:
Hi,
We have a problem about a gridftp client (uberftp) on AIX. Connention to
a
gridftp server is fine, and we can do some operations including
directory listing,
obtaining files and putting files for a while. But, It fails
unexpectedly after these
successful operations by the error
Could not list : globus_xio: Unable to connect to
C0A8:55D2:6564:3131:3831:3935:6BE5:F030:23009
globus_xio: globus_libc_getaddrinfo failed.
globus_common: Hostname and service name not provided or found
Failed to open data channel.
It seems getaddrinfo returns an IPv6 address, which cannot be
connected to, since
we do not use IPv6 in this network at all. The IPv6 address returned
by getaddrinfo
differs every time.
We are using gt-4.0.4 for an AIX client and a Linux i386 server. When we
use a
Linux i386 gridftp clients, it works very fine without any problem for
a long time.
Does someone have a clue for this problem? It seems setting
GLOBUS_XIO_TCP_SET_NO_IPV6
can force to disable IPv6 feature, but I am not sure how exactly I can
specify it in
uberftp.
Thanks,
Osamu
