Hi,
I’ve recently upgraded to 25.04 (release, not RC or beta) from the previous major version (24.10 I think? but not sure). I never installed any unstable/RC/beta versions of Fangtooth.
The upgrade completely broke any and all SMB shares for (as far as I can tell) all users.
Just to be clear: all shares worked fine before the upgrade, and has for years in some cases. Users obviously have the “SMB user” set. All have a password and are in the corresponding ACLs for the shares (and on the filesystem).
No matter which user tries, I get NT_STATUS_LOGON_FAILURE
on client side. Serverside I get an entry in the audit log of SMB mentioning NT_STATUS_NO_SUCH_USER
. Digging a bit further and googling around, I got a forum post that mentioned pdbedit -L
to check if the passwords are set, and the output was completely empty.
I checked my own user, and disabling and re-enabling the “SMB user” flag required me to reset my password, but did make my user work again for SMB. My user now also shows in the output for pdbedit -L
. So it would seem the upgrade “lost” all SMB passwords for all users.
There seems to be an RC1 forum post semi-related to this issue (post number 38009, I guess I can’t even link to internal posts for some reason, title starts with “Unable to access SMB after upgrading”).
But I don’t seem to have the exact symptoms described in the reply marked as solution. The affected users do not have a unixhash is "*"
in the output of that command, but a normal hash as expected.
Now my question is this: can I somehow force the re-creation of the samba password database? Setting it manually for every user is tedious and requires knowing the password. For some of the users I obviously don’t know the password, while for others the password is generally unknown (but set), as it’s an account for some sort of service that was just set once and not written down anywhere. I can of course in theory reset those, too, but that requires reconfiguring the services that use them (again, just tedious).
Should I just downgrade/restore backup, perform some sort of fix and re-apply the update so it can complete correctly? What’s the easiest way forward? Please note that the ZFS pools have been upgraded already (but new features aren’t actively in use yet), not sure if that’s a problem?
Thanks
Creat