Rowland Penny
2021-Dec-06 10:41 UTC
[Samba] smbstatus no longer working after upgrade to 4.13.14
On Mon, 2021-12-06 at 09:24 +0000, Julian Einwag via samba wrote:> Hi, > > we?re running a Samba CTDB cluster on Ubuntu 20.04. The cluster is > part of our windows domain. > After a recent upgrade to 4.13.14 (from 4.11.6), the smbstatus > command is no longer working. Clients can connect just fine. This is > the output produced by smbstatus -a: > > > Samba version 4.13.14-Ubuntu > PID Username Group Machine > Protocol Version Encryption Signing > ------------------------------------------------------------------- > --------------------------------------------------------------------- > db_open: Failed to initialize messaging > Failed to initialize session_global: NT_STATUS_OBJECT_NAME_COLLISION > > Service pid Machine Connected > at Encryption Signing > ------------------------------------------------------------------- > -------------------------- > db_open: Failed to initialize messaging > Failed to initialize session_global: NT_STATUS_OBJECT_NAME_COLLISION > Failed to traverse sessions: NT_STATUS_OBJECT_NAME_COLLISION > > db_open: Failed to initialize messaging > /run/samba/locking.tdb not initialised > This is normal if an SMB client has never connected to your server. > > > > > Our smb.conf: > > [global] > bind interfaces only = no > interfaces = eno1 > security = ADS > workgroup = OURWORKGROUP > realm = ****** > netbios name = ****** > csc policy = disable > winbind enum users = yes > winbind enum groups = yes > winbind use default domain = Yes > log file = /var/log/samba/log.%m > max log size = 1000 > syslog = 3 > log level = 1 auth_audit:3 auth_json_audit:3 > panic action = /usr/share/samba/panic-action %d > clustering = yes > include = registry > allow insecure wide links = yes > username map = /etc/samba/user.map > veto files = /Thumbs.db/ > min domain uid = 0 > delete veto files = yes > idmap config * : backend = tdb > idmap config * : range = 3000-7999 > > idmap config OURWORKGROUP : backend = rid > idmap config OURWORKGROUP : range = 10000-999999 > > # Template settings for login shell and home directory > template shell = /bin/bash > template homedir = /home/%U > > > ceph: config_file = /etc/ceph/ceph.conf > ceph: user_id = admin > map acl inherit = yes > store dos attributes = yesNot an expert on clustering by any means, but is ctdb running ? Also you do not seem to to have 'vfs objects = ceph' set, but you are using Samba 'ceph' parameters, have you read 'man vfs_ceph' ? Rowlad
Julian Einwag
2021-Dec-06 11:11 UTC
[Samba] smbstatus no longer working after upgrade to 4.13.14
> On 6. Dec 2021, at 11:41, Rowland Penny via samba <samba at lists.samba.org> wrote: >> : config_file = /etc/ceph/ceph.conf >> ceph: user_id = admin >> map acl inherit = yes >> store dos attributes = yes > > Not an expert on clustering by any means, but is ctdb running ? > Also you do not seem to to have 'vfs objects = ceph' set, but you are > using Samba 'ceph' parameters, have you read 'man vfs_ceph? ?You are correct, our cluster is ceph backed, though we?re not using vfs_ceph, but a kernel mount instead. Removing the parameters doesn?t change anything on our test cluster, which is running the same samba issue and experiences the same issue. We did initially test vfs_ceph, but decided not to use it. I posted a debug (smbstatus -a -d10) here: https://gist.github.com/jeinwag/e5831526749226a4c710a0c60b572ebf Looks like it's trying to communicate with ctdbd, but is failing? Part of the log: messaging_dgm_ref: messaging_dgm_get_unique returned Success messaging_dgm_ref: unique = 12399100214177679836 messaging_ctdb_ref: messaging_ctdb_init returned Device or resource busy messaging_ctdb_init failed: Device or resource busy messaging_init_internal: messaging_ctdb_ref failed: Device or resource busy msg_dgm_ref_destructor: refs=0x55f744cbc0e0 db_open: Failed to initialize messaging