nmbd fails to start with locking error on ctdb/samba upgrade

[Steve French]

The machine looks like it just got recycled (checking on that with our
testers) but the log.nmbd included the following:

[2017/02/09 18:37:31.546793,  0] ../lib/util/become_daemon.c:124(daemon_ready)
  STATUS=daemon 'nmbd' finished starting up and ready to serve connections
[2017/02/09 18:37:54.574835,  0]
../source3/nmbd/nmbd_become_lmb.c:397(become_local_master_stage2)
  *****

  Samba name server P08 is now a local master browser for workgroup
MYGROUP on subnet 192.168.1.13

  *****
[2017/02/09 18:37:54.574926,  0]
../source3/nmbd/nmbd_become_lmb.c:397(become_local_master_stage2)
  *****

  Samba name server P08 is now a local master browser for workgroup
MYGROUP on subnet 10.20.26.13

  *****
[2017/02/09 19:49:51.020635,  0] ../source3/nmbd/nmbd.c:58(terminate)
  Got SIGTERM: going down...

On Tue, Feb 14, 2017 at 12:56 AM, Amitay Isaacs <amitay at gmail.com> wrote:
> On Tue, Feb 14, 2017 at 9:06 AM, Steve French <smfrench at gmail.com> wrote:
>>
>> Looks like upgrade was from 4.5 from a few months ago to current 4.5 -
>> so shouldn't have been an issue.
>>
>> On Mon, Feb 13, 2017 at 3:55 PM, Steve French <smfrench at gmail.com> wrote:
>> > Anyone run into something like this (in upgrading to 4.5-latest
>> > ctdb/samba/nmbd)?
>> >
>> > nmbd fails to start with "invalid argument" on g_lock_tdb.2
>> >
>> > ctdb log info:
>> >
>> > 2017/02/10 23:29:37.569727 [82260]: 50.samba: Redirecting to
>> > /bin/systemctl start  nmb.service
>> > 2017/02/10 23:29:37.592666 [82260]: 50.samba: Job for nmb.service
>> > failed because the control process exited with error code. See
>> > "systemctl status nmb.service" and "journalctl -xe" for details.
>> > 2017/02/10 23:29:37.593127 [82260]: 50.samba: Failed to start nmbd
>> > 2017/02/10 23:29:37.593372 [82260]: startup event failed
>> > 2017/02/10 23:29:42.594314 [82260]:
>> > ../ctdb/server/ctdb_takeover.c:2198 Released 0 public IPs
>> > 2017/02/10 23:29:42.594355 [82260]: Running the "startup" event.
>> > 2017/02/10 23:29:42.752275 [82260]: 50.samba: Could not open tdb
>> > /var/lib/ctdb/g_lock.tdb.2: Invalid argument
>> > 2017/02/10 23:29:42.752297 [82260]: 50.samba: failed to attach to ctdb
>> > g_lock.tdb
>> > 2017/02/10 23:29:42.752305 [82260]: 50.samba: g_lock_ctx_init failed
>> > 2017/02/10 23:29:42.752309 [82260]: 50.samba: failed to attach to ctdb
>> > registry.tdb
>> > 2017/02/10 23:29:42.752443 [82260]: 50.samba: Could not open tdb
>> > /var/lib/ctdb/g_lock.tdb.2: Invalid argument
>> > 2017/02/10 23:29:42.752455 [82260]: 50.samba: failed to attach to ctdb
>> > g_lock.tdb
>> > 2017/02/10 23:29:42.752463 [82260]: 50.samba: g_lock_ctx_init failed
>> > 2017/02/10 23:29:42.752469 [82260]: 50.samba: failed to attach to ctdb
>> > registry.tdb
>> > 2017/02/10 23:29:42.752474 [82260]: 50.samba: Failed to initialize the
>> > registry: WERR_INVALID_PARAM
>> > 2017/02/10 23:29:42.752482 [82260]: 50.samba: Can't load
>> > /etc/samba/smb.conf - run testparm to debug it
>> > 2017/02/10 23:29:42.763661 [82260]: 50.samba: Redirecting to
>> > /bin/systemctl start  nmb.service
>> > 2017/02/10 23:29:42.785643 [82260]: 50.samba: Job for nmb.service
>> > failed because the control process exited with error code. See
>> > "systemctl status nmb.service" and "journalctl -xe" for details.
>> > 2017/02/10 23:29:42.786060 [82260]: 50.samba: Failed to start nmbd
>> > 2017/02/10 23:29:42.786224 [82260]: startup event failed
>> >
>> > The lockfile exists and is rather large at a size of 5206016:
>> >
>> > NR_10-23:31:16 p08 /var/log $ ls -al /var/lib/ctdb/g_lock.tdb.2
>> > -rw------- 1 root root 5206016 Feb 10 23:31 /var/lib/ctdb/g_lock.tdb.2
>> >
>> >
>> > --
>> > Thanks,
>> >
>> > Steve
>>
>
> Do you have nmbd log at higher debug?
>
> Amitay.



-- 
Thanks,

Steve