the reason why net vampire failed on one of my solaris hosts seems to be that the default python 2.4.6 that comes with solaris10 does not work correctly. after I built python 2.6.2 on that machine and rebuilt s4 I was able to vampire the domain. once that succeeded I reverted to the build using python 2.4.6 and it still seems to work as expected (changes get replicated back and forth ok). seems like net vampire either needs python >2.4.6 or that the sun provided python build is broken.
here's the error I get with python 2.4.6 trying to vampire the domain: libnet_BecomeDC() failed - NT code 0xc00020ee Traceback (most recent call last): File "/opt/samba/lib/python2.4/site-packages/samba/netcmd/__init__.py", line 99, in _run return self.run(*args, **kwargs) File "/opt/samba/lib/python2.4/site-packages/samba/netcmd/vampire.py", line 51, in run (domain_name, domain_sid) = net.vampire(domain=domain, target_dir=target_dir) RuntimeError: NT code 0xc00020ee
Well as I see it we have a couple of bugs here. First we have to see if we still support python 2.4.6 (the waf configure checks for 2.4.2), if not then a change is needed to "block" user from building something that won't work. If we should support it then we have to understand why it fails. In both case I suppose that if net vampire fails for some reason we should do the way back to remove the newly DCs.
(In reply to comment #2) > Well as I see it we have a couple of bugs here. > First we have to see if we still support python 2.4.6 (the waf configure checks > for 2.4.2), if not then a change is needed to "block" user from building > something that won't work. so far net vampire is the only thing I found that doesn't work with that python version. even if I use that version after the vampire with a newer python, the second DC works fine. > If we should support it then we have to understand why it fails. I can easily reproduce that. just let me know what data we need. > In both case I suppose that if net vampire fails for some reason we should do > the way back to remove the newly DCs. that would help very much. one would expect the master to rollback the changes in case of a failure.
We should support Python 2.4 (only Python <= 2.3 support has been dropped some months ago). I add Jelmer and hope that he might have a clue.
This has nothing to do with python versions. We somewhere hit WERR_DS_INTERNAL_FAILURE»·······W_ERROR(0x000020EE) in our code. I guess somewhere in dsdb/repl/replicated_objects.c.
(In reply to comment #5) > This has nothing to do with python versions. We somewhere hit > WERR_DS_INTERNAL_FAILURE»·······W_ERROR(0x000020EE) in our code. > > I guess somewhere in dsdb/repl/replicated_objects.c. > But how do you explain that he can reliably reproduce this when using python 2.4.6 and when he installed python 2.6.x it was ok ?
Any news?
Is this still failing? It could have been fixed in the meantime. Otherwise I close this bug.
Closing as "INVALID" for now. We can reopen it if someone still can reproduce it or has an idea how to fix it.