On 11/6/20 8:20 AM, Kelly Wang (kellythw) wrote:
Is server issue? should I change to other rhel8 server and try the same?
It does sound like some sort of server or filesystem problem, yes. The mkdir
syscall does not finish and refuses to be interrupted. I don't see any way to
modify Gnulib to work around the problem automatically. However, you should be
able to sidestep the problem with something either like this:
./configure gl_cv_func_getcwd_path_max=no
or like this:
./configure gl_cv_func_getcwd_path_max=yes
It's not clear to me which of these workarounds would be better for your system.
Didn't we have this same conversation a year ago, with GNU tar, which uses the
same Gnulib module? Are you building now on the same system you used a year ago?
https://lists.gnu.org/r/bug-tar/2019-10/msg00001.html
https://lists.gnu.org/r/bug-tar/2019-10/msg00002.html
https://lists.gnu.org/r/bug-tar/2019-10/msg00003.html
Reviewing that earlier conversation prompts me to ask: are you building on a
Samba filesystem? There seems to be a problem in this area that is reportedly
fixed by upgrading to a newer version of Samba, or by fiddling with Samba
settings (not clear). See:
https://www.cs.rug.nl/~jurjen/ApprenticesNotes/tstcg_tar.html
which is part of:
https://www.cs.rug.nl/~jurjen/ApprenticesNotes/testing_samba_to_compilation_grade.html
