OES 2018 and Samba

I have 4 instances of OES 2018, 2 are upgrades and to are new installs. The Samba install on all four is not working. I have a SR open for the upgrades awaiting back-line engineering and the developers to try and come up with a fix, but the problem is present on new installs as well.

Immediately after the install when checking the the status of SMB I see:

[2018/10/08 13:59:47.504206, 0] lib/smbldap.c:1180(smbldap_connect_system) ldap_connect_system: Failed to retrieve password from secrets.tdb
[2018/10/08 13:59:48.504391, 0] passdb/secrets.c:350(fetch_ldap_pw) fetch_ldap_pw: neither ldap secret retrieved!

When I check the status of NMB I see:

2018-10-08T13:42:19.877670-06:00 nFS04 systemd[1]: Starting LSB: Samba NetBIOS naming service over IP...
2018-10-08T13:42:19.883929-06:00 nFS04 nmb[11663]: Samba NMB daemon, /usr/sbin/nmbd is not installed. ..skipped
2018-10-08T13:42:19.884159-06:00 nFS04 systemd[1]: nmb.service: Control process exited, code=exited status=5
2018-10-08T13:42:19.884337-06:00 nFS04 systemd[1]: Failed to start LSB: Samba NetBIOS naming service over IP.
2018-10-08T13:42:19.884448-06:00 nFS04 systemd[1]: nmb.service: Unit entered failed state.
2018-10-08T13:42:19.884542-06:00 nFS04 systemd[1]: nmb.service: Failed with result 'exit-code'.

The problem I can easily see is the symlink in /usr/sbin was not placed there by the installation.

Manually placing the symlink allows "nmbd" to start but the "smbd" LDAP problem remains and I wonder what else is broken in the install (and how to fix it).

I have a hard time thinking I am the only person to see this?
  • On 08.10.2018 23:24, DanW-MHTN wrote:

    > I have a hard time thinking I am the only person to see this?
    >


    I don't honestly. Hardly anybody uses Samba on OES but CIFS. If Novell
    wouldn't have made the mistake in OES1 to try and use as much OSS code
    as remotely possible regardless of it's features and quality, Samba for
    OES wouldn't even exist. Unfortunately, they continued to drag it along
    all the time despite their own Cifs code. So now we have two SMB servers
    for OES, one of which is a Zombie.

    Note, that's not an excuse. It's there and should be supported and
    tested and working. It's just a note that it's really not so surprising
    as you think that it isn't.

    CU,
    --
    Massimo Rosen
    Micro Focus Knowledge Partner
    No emails please!
    http://www.cfc-it.de
  • On 08/10/2018 22:24, DanW-MHTN wrote:

    > I have 4 instances of OES 2018, 2 are upgrades and to are new installs.
    > The Samba install on all four is not working. I have a SR open for the
    > upgrades awaiting back-line engineering and the developers to try and
    > come up with a fix, but the problem is present on new installs as well.
    >
    > Immediately after the install when checking the the status of SMB I
    > see:
    >
    > [2018/10/08 13:59:47.504206, 0]
    > lib/smbldap.c:1180(smbldap_connect_system) ldap_connect_system: Failed
    > to retrieve password from secrets.tdb
    > [2018/10/08 13:59:48.504391, 0] passdb/secrets.c:350(fetch_ldap_pw)
    > fetch_ldap_pw: neither ldap secret retrieved!
    >
    > When I check the status of NMB I see:
    >
    > 2018-10-08T13:42:19.877670-06:00 nFS04 systemd[1]: Starting LSB: Samba
    > NetBIOS naming service over IP...
    > 2018-10-08T13:42:19.883929-06:00 nFS04 nmb[11663]: Samba NMB daemon,
    > /usr/sbin/nmbd is not installed. ..skipped
    > 2018-10-08T13:42:19.884159-06:00 nFS04 systemd[1]: nmb.service: Control
    > process exited, code=exited status=5
    > 2018-10-08T13:42:19.884337-06:00 nFS04 systemd[1]: Failed to start LSB:
    > Samba NetBIOS naming service over IP.
    > 2018-10-08T13:42:19.884448-06:00 nFS04 systemd[1]: nmb.service: Unit
    > entered failed state.
    > 2018-10-08T13:42:19.884542-06:00 nFS04 systemd[1]: nmb.service: Failed
    > with result 'exit-code'.
    >
    > The problem I can easily see is the symlink in /usr/sbin was not placed
    > there by the installation.
    >
    > Manually placing the symlink allows "nmbd" to start but the "smbd" LDAP
    > problem remains and I wonder what else is broken in the install (and how
    > to fix it).
    >
    > I have a hard time thinking I am the only person to see this?


    The novell-oes-samba package installs it's own nmbd and smbd binaries in
    /opt/novell/xad/samba/sbin/ along with suitable rc* startup scripts.

    Having just set up a minimal test server with OES2018 (Novell Samba
    eDirectory) in a new tree I can see that the Samba NMB daemon failed to
    start because /usr/sbin/nmbd is missing. However the Samba SMB daemon
    did start.

    Can you post the output from "egrep '(passdb|ldap admin)'
    /etc/samba/smb.conf" ?

    HTH.
    --
    Simon
    Micro Focus Knowledge Partner

    ------------------------------------------------------------------------
    If you find this post helpful and are logged into the web interface,
    please show your appreciation and click on the star below. Thanks.
    ------------------------------------------------------------------------
  • On 08/10/2018 22:24, DanW-MHTN wrote:

    > I have 4 instances of OES 2018, 2 are upgrades and to are new installs.
    > The Samba install on all four is not working. I have a SR open for the
    > upgrades awaiting back-line engineering and the developers to try and
    > come up with a fix, but the problem is present on new installs as well.
    >
    > Immediately after the install when checking the the status of SMB I
    > see:
    >
    > [2018/10/08 13:59:47.504206, 0]
    > lib/smbldap.c:1180(smbldap_connect_system) ldap_connect_system: Failed
    > to retrieve password from secrets.tdb
    > [2018/10/08 13:59:48.504391, 0] passdb/secrets.c:350(fetch_ldap_pw)
    > fetch_ldap_pw: neither ldap secret retrieved!
    >
    > When I check the status of NMB I see:
    >
    > 2018-10-08T13:42:19.877670-06:00 nFS04 systemd[1]: Starting LSB: Samba
    > NetBIOS naming service over IP...
    > 2018-10-08T13:42:19.883929-06:00 nFS04 nmb[11663]: Samba NMB daemon,
    > /usr/sbin/nmbd is not installed. ..skipped
    > 2018-10-08T13:42:19.884159-06:00 nFS04 systemd[1]: nmb.service: Control
    > process exited, code=exited status=5
    > 2018-10-08T13:42:19.884337-06:00 nFS04 systemd[1]: Failed to start LSB:
    > Samba NetBIOS naming service over IP.
    > 2018-10-08T13:42:19.884448-06:00 nFS04 systemd[1]: nmb.service: Unit
    > entered failed state.
    > 2018-10-08T13:42:19.884542-06:00 nFS04 systemd[1]: nmb.service: Failed
    > with result 'exit-code'.
    >
    > The problem I can easily see is the symlink in /usr/sbin was not placed
    > there by the installation.
    >
    > Manually placing the symlink allows "nmbd" to start but the "smbd" LDAP
    > problem remains and I wonder what else is broken in the install (and how
    > to fix it).
    >
    > I have a hard time thinking I am the only person to see this?


    The novell-oes-samba package installs it's own nmbd and smbd binaries in
    /opt/novell/xad/samba/sbin/ along with suitable rc* startup scripts.

    Having just set up a minimal test server with OES2018 (Novell Samba
    eDirectory) in a new tree I can see that the Samba NMB daemon failed to
    start because /usr/sbin/nmbd is missing. However the Samba SMB daemon
    did start.

    Can you post the output from "egrep '(passdb|ldap admin)'
    /etc/samba/smb.conf" ?

    HTH.
    --
    Simon
    Micro Focus Knowledge Partner

    ------------------------------------------------------------------------
    If you find this post helpful and are logged into the web interface,
    please show your appreciation and click on the star below. Thanks.
    ------------------------------------------------------------------------
  • I get the following:

    # Replaced by OES Install: " passdb backend = tdbsam"
    passdb backend = NDS_ldapsam:ldaps://192.168.1.16:636
    ldap admin dn = cn=nFS04-sambaProxy,o=mhtn
  • I have heard that before, but the DSfW documentation seems to indicate that Samba is the preferred method. Because I wanted to use DSfW a number of years ago, Samba was put in place and now... It isn't working.
  • In diving into current documentation it woudl appear that DSfW will support NSS AD which uses Novell-CIFS. So I may try uninstalling and going that route. . .

    I appreciate the gentle nudge to look deeper!
  • Novell-CIFS does not start up either. Still troubleshooting.

    I removed Samba first, then installed CIFS. It will not start. it registers with NCP, registers with GSS the errors out with "CRITICAL: BROWSER: StartupUdpPort138Server: Failed to create and bind socket. Error: 98"

    My opinion of OES 2018 not mention-able in mixed company.
  • On 10.10.2018 23:34, DanW-MHTN wrote:
    >
    > Novell-CIFS does not start up either. Still troubleshooting.
    >
    > I removed Samba first, then installed CIFS. It will not start. it
    > registers with NCP, registers with GSS the errors out with "CRITICAL:
    > BROWSER: StartupUdpPort138Server: Failed to create and bind socket.
    > Error: 98"


    Well, that error is pretty outspoken. It can't use that UDP port, most
    likely as it's occupied by something else. Check with netstat, I suspect
    Samba isn't completely gone.

    CU,
    --
    Massimo Rosen
    Micro Focus Knowledge Partner
    No emails please!
    http://www.cfc-it.de
  • I agree, one would think that asking to have novell-samba removed would actually do that. Unfortunately it did not do so. Port 138 is used by Netbios, which flagged that NMB and a couple of other samba related packages were still installed.

    I can now get CIFS to run but I lost configuration for OES (if I go into "OES Install and Configuration", all of the things that were installed and configured with the exception of LDAP are no longer shown).

    Since this was not a production server, I removed it and started over. Now I have a clean install with eDirectory, iManager NCP, NSS, NSS-AD, and Novell-CIFS. CIFS is working I think, but the NSS-AD will not connect to the domain. I will not say more now other than I a frustrated beyond civil words.
  • Hello,

    its said in the cifs-documentation of oes -server:

    11.1.11 CIFS Does Not Start If Samba is Running
    Description: CIFS server does not come up if the Samba server is running.
    Cause: CIFS cannot coexist with Samba daemons as they both use the same
    port.
    Action: Log in to the OES server as root. Use the following commands to
    stop the Samba daemons
    and restart the CIFS server:
    - rcsmb stop or systemctl stop smb.service
    - rcnmb stop or systemctl stop nmb.service
    - rcnovell-cifs start or systemctl start novell-cifs.service

    So it may be that if there has been a samba installation on this server
    before the install may be not as "clean" as it should be to make a
    working cifs installation possible.

    Regards Burkhard