[following here from comment #2 of bug #124632] @gustavoz: I've been searching for analog bugs on mailing lists, forums and the samba bugzilla, but none reported a crash due to big/low endian architectures. I think it may be due to invalid cache/tdb in /var/{cache,lib}/samba Could you please attach the logs of this operation?: 1) "/usr/sbin/nmbd -F -i -d 4 2>&1 |tee samba_nmbd.log" (root shell) 2) "/usr/sbin/smbd -F -i -d 4 2>&1 |tee samba_smbd.log" (another root shell) 3) "smbclient -L //localhost -U% | tee samba_smbclient.log" (user shell) 4) close the first two processes, if still open 5) testparm -v -s 2>&1 > samba_conf.log Thanks
Created attachment 82373 [details] testparm log
Created attachment 82374 [details] nmbd log
Created attachment 82375 [details] smbclient log
Created attachment 82376 [details] smbd log
Clearing the cache & tdb doesn't help either.
After looking into details i can reproduce this on x86 too. Add a non-existant interface to interfaces= and remote announce= a broadcast address that has no route and watch it die in about a minute after startup.
One month ago I've opened a bug in upstream buzilla [1] on this, suggesting this cause. But it shouldn't crash for this... [1] https://bugzilla.samba.org/show_bug.cgi?id=3619